Research Handbook on Human Resource Management and Disruptive Technologies 1802209239, 9781802209235

Citation preview

RESEARCH HANDBOOK ON HUMAN RESOURCE MANAGEMENT AND DISRUPTIVE TECHNOLOGIES

Research Handbook on Human Resource Management and Disruptive Technologies Edited by

Tanya Bondarouk Professor of Human Resource Management, Faculty of Behavioural, Management and Social Sciences, University of Twente, the Netherlands

Jeroen Meijerink Associate Professor of Human Resource Management, Faculty of Behavioural, Management and Social Sciences, University of Twente, the Netherlands

Cheltenham, UK • Northampton, MA, USA

© The Editors and Contributors Severally 2024 Cover image: Daniil Silantev on Unsplash. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical or photocopying, recording, or otherwise without the prior permission of the publisher. Published by Edward Elgar Publishing Limited The Lypiatts 15 Lansdown Road Cheltenham Glos GL50 2JA UK Edward Elgar Publishing, Inc. William Pratt House 9 Dewey Court Northampton Massachusetts 01060 USA A catalogue record for this book is available from the British Library Library of Congress Control Number: 2023952106

This book is available electronically in the Business subject collection http://dx.doi.org/10.4337/9781802209242

ISBN 978 1 80220 923 5 (cased) ISBN 978 1 80220 924 2 (eBook)

EEP BoX

Contents

List of figuresviii List of tablesix List of contributorsx PART I 1

INTRODUCTION Introduction to the Research Handbook on Human Resource Management and Disruptive Technologies2 Tanya Bondarouk and Jeroen Meijerink

PART II

CRITICAL PERSPECTIVES ON HRM AND DISRUPTIVE TECHNOLOGY

2

Disruptive technologies or disruptive debates? On a disrupted discussion about the future of jobs Tanya Bondarouk and Huub Ruёl

3

A self-determination theory framework to develop motivation-enhancing algorithmic management Xavier Parent-Rocheleau, Marylène Gagné and Antoine Bujold

4

Mitigating bias in AI-powered HRM 39 Melika Soleimani, James Arrowsmith, Ali Intezari and David J. Pauleen

5

Navigating through ethical dilemmas, human rights and digital governance51 Jesús Salgado-Criado and Celia Fernández-Aller

6

Algorithmic management from a ‘fault line’ to a frontline opportunity for trade unions through organizational learning pragmatist take Pierrette Howayeck

9

23

74

PART III HRM FOR DISRUPTIVE AND DISRUPTED ORGANIZATIONS 7

HRM systems and online labour platforms: survival of the (mis-)fittest?94 Anne Keegan and Jeroen Meijerink

v

vi  Research handbook on human resource management and disruptive technologies

8

9

Five decades of leadership and ‘disruptive’ technology: from e-leadership and virtual team leadership to current conversations on digital leadership Robin Bauwens and Laura Cortellazzo Human resource management and customer value in the digital economy: advancing a value co-creation perspective Jeroen Meijerink

105

120

PART IV TECHNOLOGY-DRIVEN CHANGES IN HRM PRACTICE 10

Is artificial intelligence disrupting human resource management? A bibliometric analysis Stefano Za, Alessandra Lazazzara, Emanuela Shaba and Eusebio Scornavacca

11

Engaging intentionally disconnected workers: what can HR managers in facilities with workplace personal technology bans do? Melina Bumann and Michael Wasserman

12

What decision-makers need to know about digitalised talent management166 Sharna Wiblen

13

The role of disruptive technologies in talent management in Nordic multinational enterprises 177 Violetta Khoreva, Vlad Vaiman, Tanya Bondarouk and Sari Salojärvi

14

Hiring algorithms: redefining professional roles with artificial intelligence193 Elmira van den Broek, Anastasia Sergeeva and Marleen Huysman

PART V

135

152

DIGITAL DISRUPTION OF WORK PROCESSES

15

Engagement with disruptive technology: do digital generations matter?207 Frank Stegehuis and Tanya Bondarouk

16

Artificial intelligence as a colleague: towards the workplace coexistence of people and artificial intelligence Violetta Khoreva and Katja Einola

224

17

Platform work inside organisations: an exploration of tensions in intra-organisational labour platforms Philip Rogiers, Jeroen Meijerink and Stijn Viaene

238

Contents  vii

18

Empowering or taking over? A job design perspective on the effects of cobots’ introduction in the manufacturing industry Emanuela Shaba, Alessandra Lazazzara, Luca Solari and Antonella Delle Fave

19

The accelerating disconnection of work from time and place: new questions for HR Johannes Gartner, Kristiina Mäkelä, Jennie Sumelius and Hertta Vuorenmaa

20

Keep in touch in remote workplaces: the relationship between collegial isolation and contextual work performance in remote work settings and the mediating role of relatedness Pascale Peters, Robert Jan Blomme, Martine Coun and Max Weijers

254

270

283

Index299

Figures

3.1

Differences and overlaps between AM and related constructs

24

3.2

Motivation-enhancing AM model

29

5.1

Three digital facets of organization

59

5.2

Management control system

61

8.1

Historical overview of the conversation on leadership and technology 106

9.1

Conceptual framework on HRM and value co-creation in the digital economy

123

10.1

Number of publications per year since 2003

140

10.2

Most productive authors

140

10.3

Most productive countries

141

10.4

The main keywords used in the dataset over the years

142

10.5

Authors’ keywords co-occurrence

144

10.6

Thematic map of the relationship between AI and HRM practices

145

11.1

Illustration of disruptive HR strategy of investing in worker engagement platforms for intentionally disconnected workers

158

11.2

A model of using digital workplace tools to reach intentionally disconnected workers

158

13.1

Organizational typology

186

viii

Tables

2.1

The automation probability and the share of each task of judges, an example of task-automation analysis

13

4.1

Biases in developing AI for HRM

44

4.2

Mitigation techniques

45

5.1

Levels of automation of decision and action selection

63

13.1

Overview of sample MNEs and respondents

14.1

Overview of the hiring process before and with the use of the AI tool 196

14.2

Overview of HR’s roles, shifts, and activities before and with the AI tool

197

15.1

Assumed differences between Digital Immigrants and Digital Natives in using IT

211

15.2

Interviewee function and generation

213

15.3

Stepwise visualization of analytical process

214

17.1

Organising tensions in IOLPs

243

18.1

Summary of the effect of cobot introduction on jobs according to the literature

259

20.1

Overview sample

289

20.2

Construct descriptive statistics

291

20.3

Correlations second wave and the square root of the Average Variance Extracted (in bold)

291

20.4

Direct and indirect effects

292

ix

183

Contributors

James Arrowsmith is Professor of Human Resource Management at Massey University, Auckland (New Zealand). His key areas of expertise include employee engagement, pay systems, employment regulation and flexible working time. Jim is also Co-Director of Mpower (the Massey People, Organisations, Work and Employment Research group), which aims to build better connections between the academic, policy and practitioner communities to deliver robust and relevant employment-related research. Jim is on the editorial boards of four leading academic journals and is Co-Editor in Chief of Labour and Industry: A Journal of the Social and Economic Relations of Work. Robin Bauwens is Assistant Professor of Human Resource Management at Tilburg University (The Netherlands). His research interests are situated at the crossroads of leadership, human resource management, technology and the digital transformation of work. His work has appeared in multidisciplinary journals including Journal of Business Research, Computers in Human Behavior, Journal of Leadership & Organizational Studies, and European Journal of Work and Organizational Psychology. Robert Jan Blomme is Professor of Organization Behavior at Nyenrode Business University and Professor of Management and Organization at the Open University (The Netherlands). His main research concerns psychological, sociological, humanistic and institutional aspects of organizational behaviour and organizational development. He serves as an (associate) editor and reviewer for the international academic community, including the position of Editor-in-Chief for M&O: Tijdschrift voor Management en Organisatie. Tanya Bondarouk is Professor of Human Resource Management at the University of Twente (The Netherlands). Her research focuses on the integration of human resource management (HRM) with (information) technology, which has become known as electronic HRM (e-HRM). Her research interests cover such topics as strategic e-HRM, its implementation, value creation of different e-HRM applications and international and cultural aspects of e-HRM. Her publications have appeared in journals with foci on HRM, research methodology, general management and organization studies. She is the Co-Editor of Advanced Series in Management (Emerald Publishing) and an Associate Editor of the International Journal of Human Resource Management. Antoine Bujold is a PhD Student in Organizational Behavior and Human Resources at HEC Montréal (Canada). His research focuses on the relationships between digital technologies (notably algorithmic management and artificial intelligence) and human x

Contributors  xi

behaviour in organizational settings, as well as on the impact of these technologies on human resources practices. His publications have appeared in Computers in Human Behavior Reports and Canadian Psychology/Psychologie Canadienne. Melina Bumann is Project Assistant at FIEGE Logistics (Germany). Her management and research interests focus on workplace implementation of technology, user interface, information design and service innovation. She earned her Bachelor of Media and Communications for Digital Business Degree from the University of Applied Sciences in Aachen (Germany) and her Master’s in Digital Business and Innovation Management at the Muenster University of Applied Sciences in Muenster (Germany). Laura Cortellazzo is Assistant Professor of Human Resource Management and Organization Design at the Ca’ Foscari University of Venice (Italy). She is a member of the Ca’ Foscari Competency Centre; a research centre that aims to improve individuals’ performance and employability through the development of behavioural competencies. Her research focuses on the fields of organizational behaviour, human resources management, leadership and careers. Her research work has been published in various international journals, including Journal of Vocational Behavior, Journal of Small Business Management, and Frontiers in Psychology. Martine Coun is Assistant Professor of Organization Studies and Leadership at the Open University (The Netherlands). Her research focuses on bridging leadership and HRM related topics. She is particularly interested in the role of leadership in hybrid work contexts and the consequences of remote working for collaboration in organizations for employees and supervisors. Her work has been published in The International Journal of Human Resource Management, European Management Journal, Journal of Leadership Studies, Gedrag & Organisatie and Frontiers in Psychology. Currently she serves as a Guest Editor for a special issue of Frontiers in Psychology. Antonella Delle Fave is Professor of Psychology at the Medical School, University of Milano (Italy). Her research work is centred on the study of mental health indicators, flow experience and daily experience fluctuation patterns across life domains and cultures and among individuals experiencing conditions of diversity and adversity. Together with international partners she has conducted a mixed-method design project aimed at identifying happiness and wellbeing components across countries. Her scientific production includes papers in international peer-reviewed journals and academic books. She has served as President of the International Positive Psychology Association, of the European Network of Positive Psychology and of the Società Italiana di Psicologia Positiva. She is currently Editor in Chief of the Journal of Happiness Studies. Katja Einola is Assistant Professor at Stockholm School of Economics (Sweden). Her research focuses on teams, leadership, HRM and artificial intelligence in organizations, as well as on organizational dysfunctions such as wilful ignorance. Her

xii  Research handbook on human resource management and disruptive technologies

research has been published in The Leadership Quarterly, Human Relations, Human Resource Management and Organization Studies, among others. She is fluent in six languages and has 20 years of professional experience in various leadership roles in small and large multinational firms spanning three continents. Celia Fernández-Aller is Senior Lecturer of Legal, Ethical and Social Aspects at Universidad Politécnica de Madrid (Spain). Her research focuses on the human rights approach to technology. Her research interests cover such topics as privacy and digital rights, digital divide, gender divide and ethics. She is an active member of the Sustainable Organizations Research Group and itdUPM (Human Development Technology Centre – UPM). Her publications have appeared in Science and Engineering Ethics, IEEE Technology and Society Magazine, Sustainability and DOXA Cuadernos de Filosofía del Derecho. Marylène Gagné is a John Curtin Distinguished Professor at the Future of Work Institute in the Faculty of Business and Law of Curtin University (Australia). Her research examines how organisations, through their structures, cultures, rewards, tasks and management, affect people’s motivational orientations towards their work, including volunteer work, and how quality of motivation influences performance and wellbeing in the workplace. She currently sits on several editorial boards of journals in psychology and management. She is a recipient of an American Psychological Association Dissertation Award, a Canadian Psychological Association New Researcher Award and is a Fellow of the Society for Industrial and Organizational Psychology and of the Academy of Social Sciences in Australia. Johannes Gartner is Post-Doctoral Researcher at Lund University (Sweden). His research focuses on digital entrepreneurship and technology management. His publications have appeared in journals such as Journal of Business Ethics, Technological Forecasting and Social Change, Journal of Business Research and Small Business Management. Pierrette Howayeck is a Doctoral Candidate at IAE – Sorbonne Business School and a Teaching and Research Assistant at École de Management de la Sorbonne – Université Paris I-Panthéon-Sorbonne (France). Her research focuses on the regulation of the employment relationship, particularly the analysis of algorithmic management and digital HR tools implementations, the functioning of trade unions and social dialogue. She also serves as a Graduate Editorial Intern at Socio-Economic Review, originating in the Society for the Advancement of Socio-Economics (SASE). Marleen Huysman is Professor of Knowledge and Organization at the Vrije Universiteit Amsterdam (The Netherlands). She heads the department of Knowledge, Information and Innovation (KIN) and is director of the KIN Center for Digital Innovation. Her current research interests include the reconfiguration of knowing practices in relation to the development, implementation and use of algorithmic technologies. Her research is published in books and journals such as Journal of Computer Mediated Communication, Journal of Information Technology, Journal

Contributors  xiii

of Management Studies, MIS Quarterly, Organization Science and Organization Studies. Ali Intezari is Senior Lecturer at UQ Business School, University of Queensland, Brisbane (Australia). His research interests include decision-making, human– technology interaction, wisdom theory, knowledge management and cultural studies. He has published in numerous journals such as Decision Sciences, International Journal of Information Management, Internet Technology and People, Journal of Business Ethics, Journal of Knowledge Management, Communications of the Association for Information Systems and Journal of Management Inquiry. He is the Co-Author and/or Co-Editor of the books: Wisdom, Analytics and Wicked Problems: integral decision-making in the information age, Leadership: regional and global perspectives, and Practical wisdom in the age of technology: Insights, issues and questions for a new millennium. Anne Keegan is Full Professor of Human Resource Management at University College Dublin, College of Business and Head of the HRM-Employment Relations Group (Ireland). Her research is on HRM in online labour platforms, gig work and project-based organizations. Her work has been published in peer-reviewed journals including Human Resource Management Journal, Human Resource Management, International Journal of Human Resource Management, Journal of Management Studies, Organization Studies, Journal of Applied Psychology and International Journal of Project Management where she currently sits on the Strategic Advisory Board. Violetta Khoreva is Assistant Professor of Management and Organization at Hanken School of Economics (Finland). Her research focuses on managing people and artificial intelligence in various institutes including Finnish organizations and Nordic Multinational enterprises. She has a special interest in exploring the integration of artificially intelligent colleagues into workplace ecosystems. Her research has been published in Human Resource Management, Corporate Governance: An International Review, Career Development International, Personnel Review, Employee Relations, Journal of Managerial Psychology and other outlets. She currently serves as a guest editor at Human Resource Management. Alessandra Lazazzara is Associate Professor of Organization Theory and Human Resource Management at the University of Milan (Italy). Her research interests include job crafting, e-HRM and diversity and inclusion. Her publications have appeared in HRM and OB journals such as the Journal of Vocational Behavior, The International Journal of Human Resources Management and Personnel Review. She is an Editorial Board Member of The International Journal of Human Resource Management, Industrial and Commercial Training, SN Business & Economics, Baltic Journal of Management and Prospettive in Organizzazione. She currently serves as Vice President of itAIS, the Italian Chapter of the Association for

xiv  Research handbook on human resource management and disruptive technologies

Information Systems, and as a Board Member of ASSIOA, the Association of Italian Organization Studies Academics. Kristiina Mäkelä is Professor of International Business at Aalto University School of Business and Provost of Aalto University (Finland). Her research focuses on human resource management and future of work-related issues, and in terms of her societal work, she is also involved in developing technology policy and addressing future competence needs in Finland. Her work has been published in outlets such as Journal of International Business Studies, Journal of Management Studies, Human Resource Management and Journal of World Business, among others. Jeroen Meijerink is Associate Professor of Human Resource Management at the University of Twente (The Netherlands). His research focuses on the value of human resource management in platform-based organizations, including online gig platforms (e.g. Uber and Deliveroo), social media platforms and interorganizational talent platforms (e.g. Open Opportunities). He has a special interest in the use of algorithms and artificial intelligence in human resource management. His publications have appeared in Human Resource Management Review, The International Journal of Human Resource Management, Human Resource Management and Research in Personnel and Human Resource Management, to name a few. He currently serves as Associate Editor at the International Journal of Human Resource Management and has co-edited the book Platform Economy Puzzles: A Multidisciplinary Perspective on Gig Work. Xavier Parent-Rocheleau is Assistant Professor of Human Resources Management at HEC Montreal (Canada). He holds a PhD in Human Resources Management and Organizational Behavior from the University of Quebec in Montreal and has been Visiting Scholar at Curtin University. His research is focused on algorithmic management of the workforce, workplace surveillance, quantification and datafication of work and leadership. His work has been published in influential journals of interdisciplinary fields such as Human Resource Management Review, Journal of Business Research, European Journal of Work & Organizational Psychology, Public Administration Review, Journal of Leadership & Organizational Studies, Journal of Business & Psychology and Nature Reviews Psychology. David J. Pauleen is Professor at the School of Management at Massey University (New Zealand). His current research speciality revolves around wisdom in management and, in particular, offering critical commentary on information systems research and practice. His work has appeared in numerous journals including Industrial Marketing Management, Journal of Business Ethics, Behavior and Information Technology, Communications of the AIS, Sloan Management Review and Journal of Management Information Systems. He is also Editor or Co-Editor of the books ‘Virtual Teams: Projects, Protocols and Processes’, ‘Personal Knowledge Management: Individual, Organizational and Social Perspectives’, and ‘Cross-Cultural Perspectives on Knowledge Management’, and Co-Author

Contributors  xv

of ‘Wisdom, Analytics and Wicked Problems: Integral Decision-Making in the Information Age’ and ‘Management Decision Making, Big Data and Analytics’. Pascale Peters is Full Professor of Strategic Human Resource Management at Nyenrode Business University (The Netherlands) and Visiting Professor at Inland University of Applied Sciences (Norway). She is a member of the editorial board of Tijdschrift voor Arbeidsvraagstukken and participates in Holland Management Review. She supervises PhD students and Master’s students on topics such as work-life balance, boundary management, sustainable HRM (employability, workability and vitality) and the contemporary organization of work, in particular, home-based telework, and new ways to work, currently known as hybrid working. She has published on these themes in international and national journals and books, including Human Relations, International Journal of Human Resource Management and Gedrag & Organisatie. Recently she edited the book ‘Virtual Management and the New Normal’ published by Palgrave Macmillan. Philip Rogiers is an Assistant Professor of organizational behaviour and organizational theory at the University Ramon Llull, Esade, Spain. His research focuses on the transformation and deconstruction of jobs, along with the exploration of alternative organizational forms that support a more human-centred future of work. His publications have appeared in leading journals including Academy of Management Perspectives, Academy of Management Discoveries, and the Journal of Vocational Behavior. Huub Ruël is a senior researcher affiliated with the University of Twente (The Netherlands) and Tilburg University (The Netherlands). His research focuses on the intersection of international business, international relations and ethics/morality. His most recent work investigates how corporate governance can be informed by Catholic social teaching. Besides his academic work, Dr Ruël is involved in academic development of scholars, academic programmes and curricula as well as issues of internationalization. He has published extensively in journals, via self-authored books and guest-edited books and special issues. Jesús Salgado-Criado is Senior Lecturer in the Organisational Engineering and Business Administration department in the Industrial Engineering School at Universidad Politécnica de Madrid (Spain). He is a seasoned professional with more than 30 years of experience in research, development and management in ICT companies. His area of research is digital governance in organizations and he lectures on legal, ethical and social implications of people and business analytics in several specialized schools. Moreover, he is an active member of the Sustainable Organizations Research Group and itdUPM (Human Development Technology Centre – UPM), a contributor to academic and nonacademic publications and also a participant in industry standardization bodies. Sari Salojärvi is Partner in IMS Talent Ltd and Adjunct Professor of Knowledge Capital at Tampere University (Finland). Her research focuses on the value of human

xvi  Research handbook on human resource management and disruptive technologies

capital and talent management practices in private organizations. In her main role as a business partner, she works in an executive search and leadership advisory capacity. Her special interest and research focus is on digitalization and AI of human resource management. Eusebio Scornavacca is Professor of Innovation Policy at the School for the Future of Innovation in Society, College of Global Futures and Thunderbird School of Global Management at Arizona State University (United States). His research interests include disruptive digital innovation, high-impact innovation, digital entrepreneurship, ICT for development, digital ecosystems and climate-tech. He has extensive experience developing multinational, cross-disciplinary inquiry-based learning projects. Prof. Scornavacca’s research has appeared in leading journals across multiple fields. He has held several editorial positions and served as a Track Chair at leading conferences. Anastasia Sergeeva is an Associate Professor at the KIN Center for Digital Innovation at the Vrije Universiteit Amsterdam (The Netherlands). Her research interests include technology-mediated organizational change, the transformation of professional work and the emergence of new forms of organizing due to digital technologies. She has studied these topics across diverse occupational contexts, following the introduction of such emerging technologies as surgical robotics, predictive policing and algorithmic hiring. Her work has been published in Organization Science, MIS Quarterly, and Human Resource Management Journal, among others. Emanuela Shaba is Fellow Researcher at the University of Milan (Italy). Her research focuses on the integration of work and organization of work with advancement of Industry 4.0 technology. Her research interests cover such topics as sociotechnical organizational design in light of Industry 4.0, work design and cobot implementation. She has developed a recent interest in exploring the application of algorithms in HRM-related work practices and on how virtual technologies impact work. Her publications have appeared in journals with foci in general management and organization studies, such as Business & Impresa, Studi Organizzativi and Systemic Practice, and Action Research. Luca Solari is Professor of Human Resource Management at the University of Milan (Italy). His research focuses on the interplay between individual action and organizational assets and HRM processes. His research interests focus on how to redesign organizations to promote freedom and autonomy by an active use of the potential of liberating technologies. His publications have appeared in Industrial and Corporate Change, Human Relations, Research in the Sociology of Organizations and Advances in Strategic Management. At the University of Milan, he currently serves as Director of the School of Journalism and CEO at Fondazione Unimi. Melika Soleimani holds a PhD in Management Information Systems from Massey University, Auckland (New Zealand). Her research interest is as a multidisciplinary AI generalist that thinks about developing a holistic system focusing on developing

Contributors  xvii

ethical and unbiased AI. She has published articles in academic conferences and journals such as Hawaii International Conference on System Sciences (HICSS), and International Journal of Knowledge Management (IJKM). Frank Stegehuis is a PhD Researcher at the University of Twente (The Netherlands). His research focuses on servitization from an ecosystem perspective and examines how managers can perceive and align their business ecosystem to benefit the servitization process. Jennie Sumelius is Professor of Management and Organization at Hanken School of Economics (Finland). Her research focuses on people management issues in multinational corporations, including talent management, the role of the HR function, identity work and identification and employee perceptions. Her work has been published in Journal of International Business Studies, Journal of Management Studies, Journal of World Business, Human Resource Management Journal and Human Resource Management, amongst others. Vlad Vaiman is Professor and the Associate Dean at the School of Management of California Lutheran University (United States) and a visiting professor at several premier universities around the globe. He has published seven books on managing talent in organizations and at a country level, as well as a number of academic and practitioner-oriented articles and book chapters on talent management and international HRM. His work has appeared in Academy of Management Learning and Education, Human Resource Management, International Journal of Human Resource Management, Human Resource Management Review, Journal of Business Ethics, to name a few. He is also a Founding Editor and the Chief Editorial Consultant of the European Journal of International Management, and an Editorial Board Member of several academic journals, such as European Management Review, Human Resource Management and others. Elmira van den Broek is an Assistant Professor in the House of Innovation at the Stockholm School of Economics (Sweden). Her research interest lies in the intersection of the fields of technology, work and organizations. Specifically, she explores the implications of emerging technologies such as artificial intelligence for work and organizing. Her research is primarily characterized by qualitative, ethnographic methods and a practice approach in understanding how new technologies shape knowledge work, ethics, and occupations in organizations. Her recent scholarly work has been published in MIS Quarterly and Journal of Management Inquiry. Stijn Viaene is Professor of Information Systems Management at KU Leuven (Belgium) and Professor and Partner at Vlerick Business School (Belgium). His research focuses on information systems management issues in digital innovation and transformation, business and IT alignment and business process management. He has a special interest in researching topics related to discovering what it means to create a digital enterprise in close collaboration with leaders from practice. His publications have appeared in journals such as Business and Information Systems Engineering,

xviii  Research handbook on human resource management and disruptive technologies

Technological Forecasting and Social Change, MIT Sloan Management Review, and Academy of Management Discoveries. He is the author of Digital Transformation Know How: Connecting Digital Transformation, Agility and Leadership (Acco, Leuven). Hertta Vuorenmaa is Research Program Director (Future of Work) and University Lecturer at the Department of Management at Aalto University School of Business (Finland). Her research focuses on different fringe areas of changing working life, employee engagement, wellbeing and people management (HRM) in public and private organizations. Her work has been published, for example, in Nordic Journal of Masculinity Studies and Journal of Management Studies. She is the chair of The Finnish Association of Work Life Research (FAWORE). Michael Wasserman is Professor of International Management at the Muenster University of Applied Sciences in Muenster (Germany). His research interests broadly span the strategic integration of technology and human capital in international contexts. More specifically, Mike’s work is focused on the linkages among disruptive technologies, people, sustainability and supply chains. His research has been published in a wide variety of leading journals, including the Journal of Applied Psychology, Human Performance and the Journal of Supply Chain Management. Max Weijers is a change management professional with a track record of successfully implementing organizational transformations. With 15+ years of experience in the field, he has a deep understanding of the strategies and tactics needed to drive change and improve organizational performance. In addition to change management, Max has a strong interest and expertise in workplace and workforce analytics. He has used data and analytics to inform decision-making and improve HR processes, resulting in increased efficiency and improved employee experience. As a senior manager within the Employee Experience & HR department at Capgemini Invent, he is dedicated to creating positive and engaged workplaces. Max has a proven ability to understand and address the needs of employees, resulting in improved morale, productivity and retention. Sharna Wiblen is an Assistant professor and Lecturer in Management within the Sydney Business School, University of Wollongong, Sydney (Australia) who promotes responsible talent management whereby decisions and actions are intentional, deliberate and informed. Sharna’s academic research relates to the boundaries between talent and various information and HR technologies and HR analytics with a focus on instigating, contributing to and shaping informed conversations that unpack the complexity of talent concepts, genuine talent management and future leader talent identification. Her research has featured in academic outlets such as Journal of Strategic Information Systems, The International Journal of Human Resource Management, Asia Pacific Journal of Human Resources and The Encyclopedia of Electronic HRM.

Contributors  xix

Stefano Za is Associate Professor of Organization Studies and Information Systems at University of Chieti-Pescara (Italy). His research interests include analysis and design of digital artefacts and organizational systems. He is currently focused on digital innovations and business transformation affecting people and organizations. He has published in international conference proceedings (e.g. International Conference of IS, European Conference of IS, HICSS), journals (e.g. Government Information Quarterly, Information & Management, Communication of AIS, Information Systems and eBusiness Management) and book series. He is the president of the Italian chapter of the Association for Information Systems (AIS). Moreover, he has been a member of program committees for national and international conferences in the Information Systems (IS) domain.

PART I INTRODUCTION

1. Introduction to the Research Handbook on Human Resource Management and Disruptive Technologies Tanya Bondarouk and Jeroen Meijerink

OVERALL AIM OF THE EDITED VOLUME The past decades have witnessed the emergence of many technologies – such as artificial intelligence, online platforms, internet of things and social robots, to name but a few – that have fundamentally altered human resource management (HRM), work and performance management, work-life balance, organization dynamics and discussions about the future of work. The discussions they enforce are in some ways new, but in many ways they ring a strong bell of familiarity for those who have been in HRM practice and research for a longer time. To us, the main questions remain: How should HRM research and practice respond to the observation that the technologies are evolving rapidly and aggressively, with unclear trajectories into the future, when no one in the scholarly and practitioner community can easily make an overview of existing technologies? What are the general principles that have to be developed to design technological arrangements for management of people matters in organizations? These questions confront researchers around the world, and it is the objective of this handbook to offer insights into answering them both in general and with respect to specific emerging disruptive technologies. Our objectives are to help better understand the HRM challenges posed by disruptive technologies and to develop generalizable propositions to respond to them. Another challenge posed by emerging disruptive technologies is directly linked to their hi-tech nature and the limited knowledge that most HRM researchers and practitioners have about how they work and why, and what the possible applications and consequences of their deployment are. Despite the increasing role that technology plays in how HRM and organizational actors behave, researchers in HRM and organizational behaviour still and often avoid theorizing about it (see for an overview, Landers and Marin, 2021). To this end, a variety of chapters in this edited volume explicitly address and model technology from a social science perspective. We are convinced that this approach will allow us to understand who and what is disrupted; and what disrupts it all. To facilitate this approach, we invited the contributors to this edited volume to discuss the disruptive nature of technologies and the work and organizational processes that are most influenced by entanglement with technologies. We also address critical management issues such as the ethics of these disruptions, 2

Introduction  3

digital human rights in the work environment, manipulation of digital HR information, dignity and meaningfulness of digital work, power issues and the like. This handbook focuses on current discussions and future directions for research on HRM and disruptive technologies. For a decade, the discussions have no longer centred around whether we need technologies in HRM; rather, insights are needed into how to best utilize the existing knowledge about HRM and technology and how to incorporate insights and design issues about disruptive technologies in the scholarly and practitioners’ knowledge. This handbook will combine traditional approaches with the latest insights into the role of disruptive technologies in HRM and work management.

UNPACKING THE INTERFACES OF HRM AND DISRUPTIVE TECHNOLOGIES Disruptive Technologies and Digital Disruption In this edited volume, the chapters focus on information technologies that involve computers for transmitting, storing, creating and/or exchanging information and which are associated with some kind of interruption or change in business practice. The manifestation of disruptive technologies is manifold, as is the way technology-based disruption is conceptualized in the literature. Similarly, the chapters in this edited volume focus on different types of emerging technologies that (may) have a disruptive effect. Salient examples of such technologies and the management activities they afford include artificial intelligence (Strohmeier and Piazza, 2015; Vrontis et al., 2022), online platforms (Duggan et al., 2020; Kuhn et al., 2021), algorithmic management (Kellogg et al., 2020; Lamers et al., 2022), the internet of things (Sestino et al., 2020; Strohmeier, 2020a), wearables (Garcia-Arroyo and Osca, 2021) and cobots (Habraken and Bondarouk, 2017). The academic literature and the chapters in this book alike show that the notion of disruption can be understood in different ways (Strohmeier, 2020b; Minbaeva, 2021; Møller et al., 2017). Technology-based disruption (or digital disruption) can be broadly defined as a change process that involves the use of digital artefacts (Møller et al., 2017). It is these very digital artefacts (such as artificial intelligence, platforms, etc.) that play a role in bringing about change processes and which we refer to as disruptive technologies. The plurality in definitions of disruptive technologies and digital disruption becomes salient – among others – by the difference in ‘what’ is disrupted. The new era of technological innovation, in which new generations of digital technologies are converging and undergoing widespread integration, is making whole new fields possible, including artificial intelligence, robotics and digital twins. These technologies have the potential to radically alter work life, organizational practices and social and economic institutions in organizations. Societal disruption may well be necessary and desirable for responding to pressing global problems such as climate change and energy transition. But the technologies also raise tough

4  Research handbook on human resource management and disruptive technologies

questions that are in need of HRM evaluation. A complication is they may affect the basic concepts and values that we normally appeal to in our organizational cultures and routines, such as the concept of jobs, work, workplace, responsibility and accountability. That is, digital technologies can play a role in bringing about changes in the work practices and routines of individual workers (Drost, 2023; Van den Broek et al., 2021), (HR) management processes (Minbaeva, 2021; Strohmeier, 2020b), business models (Markides and Oyon, 2010) or entire industries (Vesti et al., 2017). Equally, disruptive technologies bring about change to academic research. Digital technologies such as artificial intelligence afford novel methodologies for the empirical study of HRM, work and organizations, for instance, to analyse big datasets and/or uncover patterns of data that traditional analytical techniques do not afford. Moreover, digital techniques disrupt the conceptual basis that underpins management research (Landers and Marin, 2021; Minbaeva, 2021). Salient examples are online labour platforms, which eradicate the employment relationship as the key touchstone of HRM research (Keegan and Meijerink, 2023), or the datafication of work, which reconfigures organizational control mechanisms (Schafheitle et al., 2020). According to some, digital disruption occurs when the entity that is affected by a disruptive technology is marginalized or even completely displaced (Strohmeier, 2020b). This implies that disruptive technologies are an external factor, outside the control of the disrupted entity (e.g. an individual or organizational unit) that needs to be responded to reactively. Others suggest that organizations (or individuals) proactively operate digital technologies to bring disruption, for instance, by establishing a new venture that introduces a new digital product or market that disrupts the business of incumbent firms (Markides and Oyon, 2010; Vesti et al., 2017). As such, while digital disruptive technologies may be welcomed as opportunities for some, they may present a threat to others. Taken together, the notions of digital technologies and digital disruption have been approached in many different ways. It is not our aim to offer a unified conceptualization of these terms in this book. Instead, we seek to highlight the various ways in which digital technologies manifest and the disruptive effects they bring about. In fact, the chapters in this edited volume offer a rich overview of the many ways that disruptive technologies and HRM intersect. As we outline next, these intersections can be conceptualized along different lines, which served as the basis for the outline of the current edited volume. Human Resource Management and Digital Disruption For the purpose of this book, we define the notion of HRM broadly, along two lines. First, a portion of the chapters in this book either focus on individual HRM activities, such as hiring or job design, while others adopt a systemic perspective; the other group of chapters examine a range of HRM activities, how they appear in bundles (i.e. HRM systems) and their interrelations with disruptive technologies. Second, besides focusing on individual or bundles of HRM activities, the chapters in this edited volume also examine disruptive technologies in relation to activities and domains that are adjected to HRM, such as leadership, industrial relations, corporate

Introduction  5

governance, platform labour, work and organization psychology and business information systems. These different perspectives open the road to think about the various intersections between HRM and disruptive technologies. From the chapters included in this book, a total of four themes emerged along which this book is structured. In setting the stage, the first set of chapters (Part II) offer a conceptual basis as well as a critical reflection on HRM and disruptive technologies. Specifically, Part II includes chapters on fashions and fads in debates on artificial intelligence and the future of jobs (Bondarouk and Ruël, Chapter 2), the need and ways to overcome the dominant fatalistic investigation of the negative consequences of algorithmic management (Parent-Rocheleau et al., Chapter 3), the biases that are encoded in AI-enabled HRM and the role of HRM and AI developers in addressing them (Soleimani et al., Chapter 4), the role of digital governance in addressing the ethical dilemmas of digital technologies at work (Salgado-Craido and Fernández-Aller, Chapter 5) and the role of organizational learning in support trade unions to respond to workplace disruptions created by algorithmic management (Howayeck, Chapter 6). Part III of the book interrogates the HRM activities of disruptive and/or disrupted organizations. The chapters in this section examine how HRM and leadership activities enable organizations to respond to being disrupted by digital technologies as well as the role that HRM activities play in organizations that are considered to be disruptive (e.g. disrupting the business model of incumbent firms or industries). First, Keegan and Meijerink (Chapter 7) conceptually reflect on how HRM activities take shape in online labour platforms – such as Uber and Deliveroo, which are seen as poster-child examples of the digital disruption of the transportation industry – and specifically, how the special nature of these platforms disrupts extant thinking on fit in HRM systems. Second, Bauwens and Cortellazzo (Chapter 8), synthesize the literature on e-leadership, virtual team leadership and digital leadership. In doing so, they unpack the role of leadership in dealing with the challenges of disruptive technologies in organizations. Finally, Meijerink (Chapter 9) discusses how disruptive technologies require HRM research to examine the role of customers in value creation processes and outlines a set of HRM systems that foster value co-creation among employees and customers. Besides contributing to digital disruption and strategies to cope with disruptive technologies, HRM activities are disrupted by digital artefacts too. Accordingly, Part IV of the book includes contributions that centre on the changes brought to HRM activities by digital technologies. This section starts off with a bibliometric analysis by Za and colleagues (Chapter 10), which provides an overview of literature themes on the disruptive nature of artificial intelligence for HRM research and practice. This is followed by the work of Bumann and Wasserman (Chapter 11), which uncovers how HR managers can operate digital technologies such as chatbots to engage workers who otherwise are underserved and deprived of using digital devices at work. Third, the chapter by Wiblen (Chapter 12) provides a guide to managers that wish to use digital technologies for talent management purposes. This is complemented by the work of Khoreva et al. (Chapter 13), which offers an empirical account

6  Research handbook on human resource management and disruptive technologies

of the ways in which technological advances shape and alter talent management systems in 36 multinational corporations in Finland. Finally, van den Broek and colleagues (Chapter 14) conclude Part IV by reporting on an empirical study into the consequences of an AI hiring tool for the role of HR professionals. The book concludes in Part V with a section on the digital disruption of work processes and practices. While the other sections predominantly focus on the intersection between HRM and disruptive technologies on the organization level and technology-based change to HRM function, the final section focuses on the implications for individual employees and the way they perform their work. First, Stegehuis and Bondarouk (Chapter 15) examine the disruptive effect of digital technologies on workers and how these effects differ between end-user generations. Second, Khoreva and Einola (Chapter 16) conceptualize AI as a new organizational actor and rely on paradox theory to highlight the complex coexistence and interactions between workers and AI-as-a-colleague. Third, Rogiers and colleagues (Chapter 17) empirically uncover the tensions and disruptions that emerge when employees rely on internal labour platforms to perform part-time, fixed-term projects within the bounds of an organization, while Shaba et al. (Chapter 18) discuss the introduction of cobots at work and how these simultaneously have negative and positive effects on individual aspects of work such as autonomy, task variety and skill development opportunities. This section concludes with two studies on remote and hybrid working: Gartner et al. (Chapter 19) detail an agenda for future research that examines the impact of hybrid working for HRM activities, while Peters et al. (Chapter 20) test hypotheses on the relationship between collegial isolation and contextual work performance and the mediating role of relatedness in remote work settings.

REFERENCES Drost, J. (2023). The Irony of Rankings. PhD Thesis. University of Twente, The Netherlands. Duggan, J., U. Sherman, R. Carbery and A. McDonnell (2020). Algorithmic management and app-work in the gig economy: A research agenda for employment relations and HRM. Human Resource Management Journal, 30(1), 114–32. Garcia-Arroyo, J., and A. Osca (2021). Big data contributions to human resource management: a systematic review. The International Journal of Human Resource Management, 32(20), 4337–62. Habraken, M., and T. Bondarouk (2017). Smart industry research in the field of HRM: Resetting job design as an example of upcoming challenges. In: Electronic HRM in the smart era (pp. 221–59). Emerald Publishing Limited. Keegan, A., and J. Meijerink (2023). Dynamism and realignment in the HR architecture: Online labor platform ecosystems and the key role of contractors. Human Resource Management, 62(1), 15–29. Kellogg, K.C., M.A. Valentine and A. Christin (2020). Algorithms at work: The new contested terrain of control. Academy of Management Annals, 14(1), 366–410. Kuhn, K.M., J. Meijerink, and A. Keegan (2021). Human resource management and the gig economy: Challenges and opportunities at the intersection between organizational HR decision-makers and digital labor platforms. Research in Personnel and Human Resources Management, 39, 1–46.

Introduction  7

Lamers, L., J. Meijerink, G. Jansen and M. Boon (2022). A Capability Approach to worker dignity under Algorithmic Management. Ethics and Information Technology, 24(1), 10. Landers, R.N., and S. Marin (2021). Theory and technology in organizational psychology: A review of technology integration paradigms and their effects on the validity of theory. Annual Review of Organizational Psychology and Organizational Behavior, 8, 235–58. Markides, C.C., and D. Oyon (2010). What to do against disruptive business models (when and how to play two games at once). MIT Sloan Management Review. Minbaeva, D. (2021). Disrupted HR? Human Resource Management Review, 31(4), 100820. Møller, L., F. Gertsen, S.S. Johansen and C. Rosenstand (2017). Characterizing digital disruption in the general theory of disruptive innovation. In: ISPIM Innovation Symposium (p. 1). The International Society for Professional Innovation Management (ISPIM). Schafheitle, S., A. Weibel, I. Ebert, G. Kasper, C. Schank and U. Leicht-Deobald (2020). No stone left unturned? Toward a framework for the impact of datafication technologies on organizational control. Academy of Management Discoveries, 6(3), 455–87. Sestino, A., M.I. Prete, L. Piper and G. Guido (2020). Internet of Things and Big Data as enablers for business digitalization strategies. Technovation, 98, 102173. Strohmeier, S. (2020a). Smart HRM – a Delphi study on the application and consequences of the Internet of Things in Human Resource Management. The International Journal of Human Resource Management, 31(18), 2289–318. Strohmeier, S. (2020b) Digital human resource management: A conceptual clarification. German Journal of Human Resource Management, 34(3), 345–65. Strohmeier, S., and F. Piazza (2015). Artificial intelligence techniques in human resource management – a conceptual exploration. Intelligent Techniques in Engineering Management: Theory and Applications, 149–72. Van den Broek, E., A. Sergeeva and M. Huysman (2021). When the machine meets the expert: An ethnography of developing AI for hiring. MIS Quarterly, 45(3). Vesti, H., C. Nielsen, C.A.F. Rosenstand, M. Massaro and M. Lund (2017, December). Structured literature review of disruptive innovation theory within the digital domain. In The ISPIM Innovation Summit. Vrontis, D., M. Christofi, V. Pereira, S. Tarba, A. Makrides E. and Trichina (2022). Artificial intelligence, robotics, advanced technologies and human resource management: a systematic review. The International Journal of Human Resource Management, 33(6), 1237–66.

PART II CRITICAL PERSPECTIVES ON HRM AND DISRUPTIVE TECHNOLOGY

2. Disruptive technologies or disruptive debates? On a disrupted discussion about the future of jobs Tanya Bondarouk and Huub Ruёl

AI AND THE FUTURE OF JOBS: ANOTHER TOPIC FOR MANAGEMENT FASHION? Literally every AI-work-related discussion we’ve had with policy makers, journalists, executives and/or colleague researchers has revolved around one central question: will AI replace our jobs or how many jobs will disappear? This question could have taken a more sophisticated formulation, like: will intelligent autonomous robots replace large portions of workplaces, or will qualification requirements for the remaining workforce decrease, increase or polarize? Much conflicting and disrupting information about AI and the future of work has been thrown into the public debate. Inspired by the work of Abrahamson (1996), recently we published results of a literature review that explored electronic HRM (e-HRM) as a management fashion (Bondarouk et al., 2019), where we concluded that the appearance of e-HRM research and practices could have been attributed to management fashion forces. Based on the secondary data from multiple case studies, we saw that the e-HRM management fashion has brought several innovations in the field of HRM, and that the demand for e-HRM was shaped by sociopsychological forces that interacted with economic and technical forces (Bondarouk et al., 2019). In the model of Abrahamson (1996), groups of interrelated knowledge owners (consultants, gurus, business schools, mass media, business schools, academics) race to sense the collective preferences of managers for new tools and techniques (Clark, 2004). Management fashion setters constantly redefine their own, as well as their followers’, beliefs about which techniques lead to this progress. They deliberately produce management fashions in order to market them to fashion followers (Abrahamson, 1996). In this chapter, we do not focus on the discussion about the diversity of AI techniques available for different jobs; rather, we have observed the avalanche of discussions about the impacts of AI on jobs and work. We have seen that management and scholarly gurus created innovative, popular ideas about the (dis)appearance of jobs due to the AI invasion. These communications are often conflicting and provocative for the readership. Management consultants include significant AI producers and consumers of knowledge, and often position themselves as ‘thought leaders’ by actively creating in-house gurus. Publishers are concerned with producing and 9

10  Research handbook on human resource management and disruptive technologies

distributing ideas about AI and the future of jobs that are likely to have mass appeal (Clark, 2004). Two reasons for our interest in management fashion and the discussion about the disruptive impact of AI and the future of jobs are concerned with the question of whether academic knowledge is developing independently from fashion setters, and with our wish to understand the success and impact of this fashion on the direction of the debate. We start with a brief overview of the very rapid flood of debates about an AI-enabled future of jobs. In our observation, the recent debates took off in 2013, but already within less than a decade we can reflect on the influence of those contests.

POLARIZED DEBATES ABOUT AN AI-ENABLED FUTURE OF JOBS Without pretending to present all those reports, we provide below a brief illustration of what we think is an exemplary representation of a variety of forecasts published up to now. In 2013, the famous research paper from the Oxford Martin School discussed the question of how susceptible jobs were to technological developments (Frey and Osborne, 2013). As with every era of technological change, some jobs were seen to be replaced by machines, while new jobs were seen as newly created and existing jobs would take on new and different tasks. Although experts said that it was difficult to predict exactly which jobs would be affected, Frey and Osborne (2013) claimed to have a greater clarity about the types of jobs automation would affect in the near future, thus shedding light on which categories of workers were at greatest risk. They predicted that 47% of the US workforce was at high risk of automation and that, after several years of decline in middle-income jobs, there was a wider range of skill levels that were likely to be affected. In particular, low-skilled jobs beyond some of the already automated routine tasks were viewed as being at risk. In our view, the paper by Frey and Osborne (2013) deserves more attention. The authors labelled 70 of the 702 jobs from the O*NET1 online job database manually as either ‘automatable’ or ‘not automatable’. This labelling, as the authors admitted, was a subjective assignment, based on ‘eyeballing’ the job descriptions from O*NET. Labels were only assigned to jobs where the whole job was considered (non)automatable and to the jobs that their workshop participants were the most confident about. The paper gave raise to numerous debates. The methods were used to calculate the impact on jobs in other countries, including Germany, Finland and Norway (Bonin et al., 2015; Pajarinen et al., 2015). This was usually done by matching each job from O*NET to the locally used standardized name. Due to differences in the economies, a different percentage of the workforce was modelled to be affected by this change; for example, only one-third in Finland and Norway were predicted to be at risk compared to 47% in the US (Brandes and Wattenhofer, 2016). The manuscript of Frey and Osborne (2013) pulled off a magic trick: by showing opaque results it produced great disruption itself by dividing the reading audience

Disruptive technologies or disruptive debates?  11

into a party that believed the ‘magic’ computerization percentages and another that had doubts. The seminal work of Frey and Osborne (2013) did an excellent job of starting the discussion going. In this way, the management idea was set off as a fashion. For example, Smith and Anderson (2014) reported that among the 1896 respondents to their survey, two-thirds of the respondents anticipated that robotics and artificial intelligence would permeate wide segments of daily life by 2025, with huge implications for a range of industries such as health care, transport and logistics, customer service and home maintenance. But even as respondents were largely consistent in their predictions for the evolution of technology itself, they were deeply divided on how advances in AI and robotics would impact the economic and employment picture over the decade of 2020–30. The vast majority of respondents expected AI to do much of the work of humans in approximately 50 years, while four out of five among those same respondents said, ‘but not my job’. It is an interesting mathematical paradox: 66% of respondents believed that AI would change jobs, but 80% said it wouldn’t change their jobs. Some scholars present this as a typical example of overvaluing one’s own expertise and skills, while undervaluing the skills of others (e.g. Eubanks, 2019). Let us take other examples. The Dutch report of the Rathenau Institute that came in 2015 (Van Est and Kool, 2015), in the assignment of the Dutch Ministry of Social Affairs and Employment, documented that technological developments were undergoing a new phase, and business, society and people were confronted with new technological possibilities like artificial intelligence and robots, self-driving cars, censored networks, digital cameras, lab-on-a-chip and commercial drones (Van Est and Kool, 2015, p. 16). Speaking of AI, interconnectedness was viewed in general as a crucial success factor. The authors claimed that, due to AI development, the world would develop complex connections of supply chains, from designers to manufacturers, from distributors to importers, wholesalers and retailers, which would allow billions of products to be made, shipped, bought and enjoyed in all corners of the world. In the same year, 2015, the McKinsey Global Institute published their report about digital platforms and how they would inject new momentum into job markets. There is a ‘stubborn disconnect’ between people and jobs, reads the report (Manyika et al., 2015): about 30%–45% of the global working-age population were unemployed, inactive or working part-time while, it claimed, by 2025 digital platforms would account for 2% of global GDP and increase employment by 72 million FTE positions. This report predicted that, in particular, the time required for searching and offering jobs would be shortened – by 2025, 200 million who were inactive or working part-time in 2015 could work additional hours through such platforms in the future. Similar trends were shown in the report of the Institute for Public Policy Research (Dolphin, 2015). Although the report sounded optimistic in general about the effect of technological innovations in boosting productivity and creating prosperity for all, the authors of various essays in the report could not but worry that the gains would not be widespread. A careful reader clearly finds concerns that no guarantee was projected that there would be enough new jobs to offset the losses caused by AI devel-

12  Research handbook on human resource management and disruptive technologies

opments. This report suggested that productivity gains from technological innovation would increasingly accumulate to the ‘owners of technology and the (relatively few) workers required to operate it, while the vast majority would face stagnant real wages at best, unemployment at worst’ (Dolphin, 2015, p. 13). In this report a call was made for new skills, which were presented as a big part of the solution: digital literacy, entrepreneurial, scientific, creative and emotional skills. The overall message was clear: economists thought that technological innovation would cause major disruption to the European labour market during the next decade (Dolphin (2015). However, the report failed to articulate the precise path of the change and left the reader in confusion as to whether the likely results would be an increased polarization of the workforce and a tendency towards greater inequalities of income and wealth. Moving forward in time, the report of RAET HR Benchmark (RAET, 2016) claimed that AI would change the work of SMEs – they would become micro-multinationals, while start-ups would be ‘born global’. Indeed, Amazon hosted more than 2 million third-party sellers, and 10 million small businesses have become merchants on Alibaba platforms (RAET, 2016). An interesting observation regarding this report is that its attention was turned to the management processes within the AI-enabled environment. Thus, this research among almost 2000 respondents found that 70% of managers expected changes in competences of the workforce due to the AI disturbance, and 75% of employees without managerial responsibilities expressed that, with new technologies, they hoped to be able to steer their own careers. At the same time, only 30% of the respondents from the research reported that they ‘took the steering wheel of their own careers in their hands’ (RAET, 2016, p. 12). With this observation, we see yet another example of strong claims for AI-enabled effects, while distantiating from these claims when it comes to one’s own experience. In 2016, the World Economic Forum (WEF) published results of an extensive survey of chief human resources and chief strategy officers of global employers (371 companies representing more than 13 million employees), providing insights into the future of jobs (World Economic Forum, 2016). In the WEF report, AI was rated lower as the driver for the changes in jobs (only 7%) in comparison to cloud technology (34%), geopolitical volatility (21%), or emerging markets (23%). This research (published three years after the paper of Frey and Osborne, 2013) concluded that the accelerating pace of technological and AI disruption should be considered in integration with demographic, geopolitical and socioeconomic disruptions. Research attention – the report reads – should be given to explore the cumulative effects of all interruptions, if possible (World Economic Forum, 2016). These interruptions were viewed as ‘transforming industries and business models, changing the skills that employees need and shortening the shelf-life of employees’ existing skill sets in the process’ (ibid., p. 20). For example, rather than completely replacing existing occupations and job categories, technological disruptions such as AI were considered to be likely to substitute for specific tasks. The prevalence of insufficient understanding of disruptive changes in integration was viewed to be the main barrier to managing change and to explain the mismatch between the magnitude of the upcoming changes

Disruptive technologies or disruptive debates?  13

Table 2.1

The automation probability and the share of each task of judges, an example of task-automation analysis

Task Description

p

Share

Write decisions on cases

1

5.1

Instruct juries on applicable laws, direct juries to deduce the facts from the evidence presented

1

3.4

Monitor proceedings to ensure that all applicable rules and procedures are followed

1

8.0

Rule on admissibility of evidence and methods of conducting testimony

0.94

5.3

Preside over hearings and listen to allegations made by plaintiffs to determine whether the

0.46

5.9

and hear their verdicts

evidence supports the charges Read documents on pleadings and motions to ascertain facts and issues

0

10.1

Settle disputes between opposing attorneys

0

4.6

Participate in judicial tribunals to help resolve disputes

0

6.6

Award compensation for damages to litigants in civil cases in relation to findings by juries or by 0

3.8

the court Supervise other judges, court officers and the court’s administrative staff

Notes: Source:

0

8.5

p = given probability to be automated; Share = how much time is spent to perform the task. Adapted from Brandes and Wattenhofer (2016).

and the relatively fearful actions taken by companies to address these challenges so far (ibid., p. 36). In 2016, a call was made for a critical assessment of the claims made by the paper of Frey and Osborne (2013). Indeed, as Frey and Osborne (2013) suggested, if we know that a job is 100% automatable, we also know that every task of that job must be completely automatable. But what if a job is 87% automatable? Is every task 87% automatable? Or are 87% of the tasks completely automatable, and 13% not at all? These questions were raised by Brandes and Wattenhofer (2016), who called for deeper studies, by looking not at jobs but at the tasks that make up a job. Brandes and Wattenhofer (2016) continued the exploration of Frey and Osborne (2013) by offering a nuanced model that moved the discussion away from ‘job’ towards ‘task’ calculation, and from the frequencies with which a task is performed towards task shares – time that is spent doing this task. From numerous job analysis examples in the study of Brandes and Wattenhofer (2016) we pick up only one – the job of judges, which was assigned an automation probability of 40% in Frey and Osborne (2013). A sample of tasks, their probabilities and their shares are shown in Table 2.1. The tasks within the job of judge that could be automated were grouped in two sets: preliminary hearings, which includes making first assessments, and ensuring that the procedures in court are followed. The tasks that involve sentencing (or the preparation thereof) have been assigned low automation probabilities. The race to find the Holy Grail of AI-future-of-jobs links has continued; however, although many organizations have begun to adopt AI, the pace and extent of adoption has been uneven. Nearly half of respondents in yet another McKinsey survey on AI adoption said their companies have embedded at least one AI capability in their business processes and another 30% were piloting AI. Still, only 21% reported their

14  Research handbook on human resource management and disruptive technologies

organizations have embedded AI in several parts of the business and barely 3% of large firms have integrated AI across their full enterprise workflows (McKinsey Global Institute, 2018). The same report showed that highly digitized companies tended to invest more in AI in the hope of gaining greater value from its use. At the sector level, the gap between digitized early adopters and others was clearly widening: sectors highly ranked in the Digitization Index (DiGiX), such as high-tech and telecommunications, and financial services were shown to be leading in AI adoption and as having the most ambitious AI investment plans, while laggards found it harder to catch up. As the report of McKinsey (2018) suggested, many companies and sectors lagged behind in AI adoption: not developing an AI strategy with clearly defined benefits or finding talent with the appropriate skill sets and lacking ownership and commitment to AI on the part of leaders. Organizations were falling short in robust data capture and governance processes as well as modern digital capabilities, and reported difficulties in building or accessing the requisite infrastructure. An important observation reported by McKinsey (2018) was that demand for AI-related skill was far outperforming the supply; and competition for workers who possess those skills was fierce. The Economist sealed the discussion in May 2019 with the argument that the world was ‘enjoying an unprecedented jobs boom’ (The Economist, 2019). Their report claimed that the unemployment rate in America was 3.6%, the lowest in half a century; and two-thirds of the members of the OECD enjoyed record-high employment among 15- to 64-year-olds. While in France, Spain and Italy, where joblessness was still relatively high, working-age employment was already close to or exceeded 2005 levels. These results should be considered with a high degree of consciousness that the COVID-19 pandemic has dramatically changed the employment picture since 2019. However, for our discussion, it is interesting to see that the issue of a jobless future gradually got replaced ‘by a series of complaints about the quality and direction of work. These were less tangible and harder to judge than employment statistics’ (The Economist, 2019). If we are to summarize this historically short rise and fall of the debate about AI and the future of jobs in the popular literature, we would emphasize that many claims were made to shed light on which categories of workers were at greatest risk due to the AI disturbances. Participants in the large-scale surveys showed certain beliefs that AI would change jobs, but not their own jobs. The job market showed a so-called stubborn disconnect between people and jobs (Manyika et al., 2015) – between the unemployed working-age population and growing demand for highly skilled workers. Many reports plead for an examination of the cumulative effects of integrations of interruptions including but not limited to AI, as the prevalence of insufficient understanding of disruptive changes in integration was viewed to be the main barrier to manage the change.

Disruptive technologies or disruptive debates?  15

FROM OPPOSITE VIEWS ON THE IMPACT OF AI – TOWARDS NEW RESEARCH QUESTIONS Having reviewed the opinions of management fashion setters, we turn our attention to the scholarly debates. Scholars have actively joined this polarized debate. In the classic debate regarding the relation between technological change and employment, two opposing visions predominated. On the one hand there was the ‘upward spiral’: technological innovation gives rise to higher labour productivity, which in turn generates lower manufacturing costs, cheaper products, increased purchasing power, a growing market and eventually more jobs. In the second scenario, technological innovation also gives rise to higher labour productivity. Here, however, this is seen as resulting in a decrease in jobs, as labour is widely replaced by technology, causing a decrease in purchasing power, lower consumption and a shrinking market; hence a ‘downward spiral’ (Van Est and Kool, 2015; Kool et al., 2015). The presence of two conflicting research camps around the subject of AI and the future of jobs was further fuelled by the crumbling of an existing consensus: the traditional consensus that technological growth is at the expense of jobs in the short term but rapidly creates new jobs – in one to two years – has been falling apart since 2013, powered by management fashion setters (Van der Zee, 2015). Researchers belonging to the stream questioning the stated consensus (i.e. downward spiral) include, for instance, Brynjolfsson and McAfee (2014) and Ford (2015). These authors argue that machines, more than in the past, would replace people; according to them, the balance between job creation and job loss has shifted to the latter. Brynlofsson and McAfee (2014) highlighted the inability of people’s skills and organizations to keep pace with technical change, resulting in the appearance of what these authors term ‘the great decoupling’. In other words, they saw a continuing trend of increasing labour productivity but a drop in labour demand. Despite the above, Brynlofsson and McAfee (2014) suggested that a doomsday scenario was preventable if businesses started competing using machines instead of against them; they proposed that humans were simply not being creative enough (Brynlofsson and McAfee, 2014; Bernstein and Raman, 2015). Ford (2015), on the other hand, argued that robots and AI were going to consume much of the base of the job skills pyramid while, in addition, the top tier would not remain a safe harbour due to developments in AI applications. Therefore, a larger number of people would be fighting for an ever smaller number of jobs unless a guaranteed basic income was realized. Adding to the pessimistic side of future employment were the many replication studies of the work conducted by Frey and Osborne (2013) (e.g. Baert and Ledent, 2015; Pajarinen and Rouvinen, 2014; Schattorie et al., 2014). Unlike those viewing the future gloomily, Miller and Atkinson (2013) presented a brighter perspective (i.e. upward spiral). They stated that the pessimists assumed a completely wrong link between technological change and employment. The main reason Miller and Atkinson (2013) provided for the position that robots would not leave humans massively unemployed was that human wishes were close to infinite and, hence, as long as that was the case, there would be a continuing need for labour.

16  Research handbook on human resource management and disruptive technologies

Another positive mind was Bainbridge (2015), who, besides the argument that technology can also create jobs and bring down barriers of entry, stated that people increasingly provided the competitive edge as ‘competition lies in the quality of service that only people can deliver because people are prepared to pay a little more for quality service and positive interaction’ (Bainbridge, 2015, p. 81). Davenport and Kirby (2015) likewise offered a less grim outlook as, in their view, human work can flourish when automation is be reframed as augmentation, which ‘means starting with what humans do today and figuring out how that work could be deepened rather than diminished by a greater use of machines’ (Davenport and Kirby, 2015, p. 60). To some extent this view fits with Brynlofsson and McAfee’s (2015) statement that humans were not being creative enough. Schouteten (2015) highlighted the fact that technology in itself does not determine the function structure of work, and hence, employment; rather, it is the combination and alignment with organizational design principles or organizational choice (i.e. the space/freedom to make decisions regarding the organization of work). Having described two polar research camps on the view of the impact of AI on jobs, it is not difficult to observe that these debates can form an endless curve, and that these two camps would never agree with each other. Therefore, we suggest calling a halt to this debate, as there exist contra-arguments enough to fuel it from both sides. Every scholar should take his/her responsibility for the arguments in this debate, as do we in this chapter. Although we acknowledge that (certain) jobs might drastically change, we do not subscribe to the outlook of a jobless future. AI is not (yet) capable of outperforming humans in every aspect (Bernstein and Raman, 2015). Even were this not the case, as well as for the aspects in which AI already outperforms humans, there remains freedom in how to implement these developments. Of course, the power of AI should not be underestimated. For example, the Institute for Public Policy Research described that the amount and richness of data traversing global networks generated by the digital economy in 2017 alone exceeded the total amount of data accumulated between 1984 and 2012 (Dolphin, 2015). Eubanks (2019) provides another example: when Microsoft turned the full power of its web services towards translating Wikipedia’s three billion words across five million articles from English to another language at its Ignite event in 2016, the translation occurred in less than 1 second. However, despite that translation and all the power at their disposal AI systems ‘cannot call you on a phone and have an hour of a spontaneous jolly conversation with you’ (Eubanks, 2019, p.30). There is still a huge gap between the capabilities of AI today and the very human nature of life and work. Some scholars claim that the quality of communication and specific team and companionship skills gains even greater importance with the new technologies entering the workplace (De Graaf, 2015): social learning and imitation, emotions, learning, social competences, maintaining social relationships and exhibiting distinctive personality (De Graaf, 2015; Multu, 2011). Paraphrasing Schouteten (2015), AI in itself does not determine the future, but the decisions humans make regarding them.

Disruptive technologies or disruptive debates?  17

We recommend building on the concept of choice in order to stop viewing technological developments as a cause of diminishing jobs and to start seeing them as an opportunity. We borrow an idea from Davenport and Kirby (2015) to suggest reframing the threat of an AI impact with an opportunity for augmentation. Instead of asking a traditional question: What and how much of tasks will be taken over by AI? We should probably ask: ‘What new feats might people achieve if they had better-thinking AI to assist them’? While AI, with its efficiency-minded enterprises, deploys computers to chip away job tasks that can be codified, augmentation starts with understanding what people do, how they perform their tasks and how these tasks and achievements can be deepened with the help of computers. With this new mindset, workers will be able to see AI as a partner with which to collaborate in creative problem solving. Some new insights should become clear to employers: it is time to acknowledge that the combination and integration of humans and computers is better than either working alone (inspired by Davenport and Kirby, 2015). But let’s advance this discussion further. In the end, in a combination of jobs and AI, in many cases, the conversation boils down to productivity, impact and value creation. Instead of endlessly debating whether and/or which jobs will be replaced or augmented, we need to ask: Will AI help us to be more productive? To improve business and society? Popular sources estimate the costs of productivity lost due to non-job related administrative tasks at $4600 per employee annually (Gorman, 2017). If we transfer these costs to a company with 500 employees, we need to assume that that medium sized company will spend on average about $2.3 million every year on tasks that are not directly linked to its primary processes. Then another question arises: How can we delegate as many of such tasks as possible to AI? And if we are able to do it, how can we ensure that the workers devote more time and effort to creating value for the business, customers, clients and society? As a sub-conclusion, we would summarize our overview of what scholars write about polarization of jobs due to the latest technological developments (Brynjolfsson and McAfee, 2014), and expect that jobs in which humans have advantages over computers will increase in number, while jobs where computers and robotics have advantages will decrease (Levy and Murnane, 2013). However, we suggest that following one of these two debate camps is futureless for science and for practice. In our view, we need to focus on questions like: How can we delegate to AI tasks that are not related to primary work processes and do not add value to businesses? And how can we change human mindsets to see AI as a supportive tool instead of fearing it? Based on the discussion by Levy and Murnane (2014), we can expect that, while job numbers will change, AI will also change the nature of those jobs, requiring workers with new skills involving problem-solving and communication – things that are difficult for computers to match. Now, we will turn our discussion to the changing nature of work.

18  Research handbook on human resource management and disruptive technologies

BEYOND THE DEBATE ABOUT AI AND THE FUTURE OF WORK: MACRO-LEVEL FACTORS If we summarize our observations from the literature review, we see an equal number of reasons to be hopeful and to be concerned. Reasons to be hopeful about the future of work with AI are based on four observations. First, advances in technology may displace certain types of work, but historically they have been a net creator of jobs. Second, workers will adapt to these changes by inventing entirely new types of work and by taking advantage of uniquely human capabilities. Third, AI will hopefully free us from day-to-day drudgery, such as administrative tasks, and allow us to define our relationship with ‘work’ in a more positive and socially beneficial way. Finally, we as a society will remain in control of our own destiny through the choices we make. Reasons to be concerned cover three observations. First, impacts from automation have thus far impacted mostly manual/physical employment, but the current and expected wave of AI-based innovation threatens to upend white-collar work as well. Second, certain highly skilled workers will succeed wildly in this new environment – but far more may be displaced into lower-paying service industry jobs at best, or permanent unemployment at worst. Finally, our educational system is not yet adequately preparing us for the work of the future and our political and economic institutions are poorly equipped to handle the hard choices involved in dealing with these challenges. However, there is a set of issues that need our attention. First, the pace at and extent to which AI will be adopted and impact actual jobs will also depend on labour market dynamics – including labour supply quantity and associated wages – and labour market dynamics may result in different outcomes in developed and developing economies. For example, in quite a number of developed economies, AI innovations may help to reduce labour market shortages (as we write in 2022), whereas in developing economies, AI innovations may cause higher unemployment and increase the mismatch between labour supply and demand. Furthermore, the benefits for businesses that go beyond AI labour substitution and augmentation need to be considered more in depth. For example, will the adoption of an AI innovation that replaces jobs also contribute to a better consumer experience, to improved quality for the customer, and will an AI innovation contribute to more sustainable business practices? Does the implementation of AI bring a long-awaited competitive advantage or has it already become a competitive necessity? Finally, social norms, social acceptance and various regulatory factors are affecting the timing of AI innovation adoption in organizations and in societies. For example, how do societies respond to AI-based healthcare innovations and disruptions and, consequently, jobs in the healthcare sector? What if patients could enter a hospital without even seeing a human – being screened and assessed from A to Z and receiving a fully AI innovation-based medical intervention? Would patients and societies accept this? Such questions can be transferred to any type of industry. All these factors need continuous and very nuanced consideration, since with every new step they will re-enter the debate.

Disruptive technologies or disruptive debates?  19

CONCLUSION AND RECOMMENDATIONS FOR ACTION For businesses to capitalize on new opportunities, they will need to put talent development and future (oriented) workforce strategy front and centre to their growth. Firms can no longer be passive consumers of ready-made human capital. Firms need a new mindset to satisfy their talent needs and to sustain and improve business performance in financial terms, but also in terms of stakeholder expectations. This entails several major changes in how business views and manages talent, both in the short term as well as in the longer term. Companies considering the option of building their own AI solutions will need to consider whether they have the capacity and the business climate to attract and retain workers with these specialized skills. Business leaders will need to manage skills disruption as an urgent concern. They must understand that talent is no longer a long-term issue that can be solved with tested conventional approaches that were successful in the past or by instantly replacing existing workers. Instead, as the rate of skills change accelerates across both old and new roles, proactive and innovative skill-building and talent management are urgent issues. Offering workers jobs that can be performed anytime, anyplace and with different types of contracts will require an HR function that is rapidly becoming more intelligent – one that employs new kinds of analytical tools to spot talent trends and skills gaps and provides insights that can help organizations align their business, innovation and talent management strategies, to outperform machine intelligence in humanistic value-based choices. HR functions will need to build a new approach to workforce planning and talent management, where better forecasting data and planning metrics will need to be central. Earlier mapping of emerging job categories, anticipated redundancies and changing skills requirements in response to the changing environment will allow businesses to form effective talent planning. As physical and organizational boundaries have blurred in many circumstances, organizations will need to become significantly more agile in the way they think about managing people’s work and about the workforce as a whole. Work is what people do and not where/when they do it. The COVID-19 pandemic has made this clear to more organizations and business leaders. Businesses will increasingly connect and collaborate remotely with freelancers and independent professionals through digital talent platforms. Modern forms of association such as digital freelancers’ unions and updated labour market regulations will increasingly begin to emerge to complement these new organizational models. For policymakers, an important set of regulations concerns the portability of safeguards and benefits between jobs and the equivalent treatment in law of different forms of labour and employment types. Following our optimism about the future of jobs, we nevertheless assume that, even if there will be enough work for people in the coming decades, the transitions that will accompany AI adoption will take significant effort. Minimizing administrative tasks and enabling workers to perform meaningful work will require a new level of emotional and social intelligence amongst business leaders. Furthermore, the workforce will also need higher cognitive skills, particularly critical thinking, creativity and complex information judgement and assessment processes.

20  Research handbook on human resource management and disruptive technologies

AI’s impact on work, in our view, will be both profound in particular labour segments as well as rather limited in others. Some occupations, as well as the demand for some categories of skills, will decline, while others will grow and change as people work alongside ever-evolving and increasingly capable technologies. Despite the progress, many complex problems remain that will require more scientific breakthroughs. The harder issues are in what is usually referred to as ‘artificial general intelligence’, where the challenge is to develop AI that can tackle general problems in much the same way that humans can. Many researchers consider this to be decades away from becoming reality.

EPILOGUE Writing this chapter made us aware of two observations that are relevant for future scholarly work on the role of technological developments and the future of jobs. The first one is that scholars need to be aware of their responsibilities and the impact of their claims and findings on societal debates. Bold, un-nuanced statements will not be helpful to advance societal debates. In the AI and the future of jobs debate, many such bold statements supported by empirical findings have been made in this debate. All those research projects deserve great credit. At the same time, we – scholars – have to anticipate that societal debates may divorce from nuances that are usually very important in the empirical findings (time, location, contexts, industry, etc.). Scholarly work must always be nuanced, since it is the nature of scholarly work. In the case of the AI and the future of jobs debate doomsday scenario-like claims have not been very helpful for societies to understand what and how AI innovation disruption will affect the future of jobs. They have scared off policy makers, HRM professionals, politicians and societies at large unnecessarily. Second, and partially in support of the previous remark, the expected jobless future caused by AI disruption seems not arise on the horizon yet. Many developed economies face severe labour market shortages in hospitality, healthcare, education, agriculture, home delivery, construction and infrastructure building work to mention a few; and despite predicted economic slowdowns due to international relations uncertainties, the high demand for people in these industries is expected to stay. AI-based innovations in many cases are highly needed to help to solve these labour market shortages, and not to make people worry that they may lose their jobs, but to enable and to serve the current workforce and help organizations to sustain, and stay in, business.

NOTE 1.

O*NET is an application that was created for the general public to provide broad access to the O*NET database of occupational information. The site is maintained by the National

Disruptive technologies or disruptive debates?  21

Center for O*NET Development, on behalf of the US Department of Labor, Employment and Training Administration (USDOL/ETA); see https://​www​.onetonline​.org/​.

REFERENCES Abrahamson, E. (1996). Management fashion. Academy of Management Review, 21(1), 254–85. Baert, A., and P. Ledent (2015). De technologische revolutie in België. Economic Research. ING. Bainbridge, S. (2015). In the future, what will people do? In T. Dolphin (Eds.), Technology, Globalisation and the Future of Work in Europe: Essays on Employment in a Digitised Economy (pp. 80–85). London: Institute for Public Policy Research (IPPR). Retrieved on 12.08.2022 from http://​www​.ippr​.org/​publications/​technology​-globalisation​-and​-the​-future​ -of​-work​-in​-europe. Bernstein, A., and A. Raman, (2015). The great decoupling. Harvard Business Review, 93(6), 66–74. Bondarouk, T., H.J.M. Ruël and B. Roeleveld (2019). Exploring Electronic HRM: Management fashion of fad? In: A. Wilkinson, N. Bacon., S. Snell, and D. Lepak (Eds.), The SAGE Handbook of Human Resource Management, Second Edition (pp. 271–91). SAGE Publications, London. Bonin, H., T. Gregory, and U. Zierahn (2015). übertragung der Studie von Frey/Osborne (2013) auf Deutschland (No. 57). ZEW Kurzexpertise. Brandes, P., and R. Wattenhofer (2016). Opening the Frey/Osborne black box: Which tasks of a job are susceptible to computerization? arXiv preprint arXiv:1604.08823. Brynjolfsson, E., and A.McAfee (2014). The Second Machine Age. Work, Progress, and Prosperity in a Time of Brilliant Technologies. New York: W.W. Norton & Company. Clark, T. (2004). Strategy viewed from a management fashion perspective. European Management Review, 1(1), 105–11. Davenport, T., and Kirby, J. (2015). Beyond automation. Strategies for remaining gainfully employed in an era of very smart machines. Harvard Business Review, June 2015, 58–65. De Graaf, M. (2015). Living with Robots. Investigating the User Acceptance of Social Robots in Domestic Environments. Enschede: Center for Telematics and Information Technology, University of Twente. Dolphin, T. (Ed.) (2015). Technology, Globalisation and the Future of Work in Europe: Essays on Employment in a Digitised Economy, IPPR. Retrieved on 13.07.2022 from http://​ www​.ippr​.org/​publications/​technology​-globalisation​-and​-the​-future​-of​-work​-in​-europe. Economist, The (23 May, 2019). The rich world is enjoying an unprecedented jobs boom. The Economist Group Limited, London 2019. Eubanks, B. (2019). Artificial Intelligence for HR. Kogan Page Ltd., London. Ford, M. (2015). The Rise of the Robots: Technology and the Threat of a Jobless Future. Basic Books. Frey, C.B., and M.A. Osborne (2013). The Future of Employment: How Susceptible are Jobs to Computerization? Oxford Martin School Publications. Retrieved from http://​www​ .oxfordmartin​.ox​.ac​.uk/​downloads/​academic/​The​_Future​_of​_Employment​.pdf. Gorman, C. (2017). TLNT Talent Management and HR. Retrieved on 22.08.2022 from https://​ www​.tlnt​.com/​poor​-admin​-is​-costing​-you​-4600​-per​-employee/​. Kool, L., R. Van Est, I. Van Keulen, and A. Van Waes (2015). Inleiding. In R. Van Est and L. Kool (Eds.), Werken aan de robotsamenleving: Visies en inzichten uit de wetenschap over de relatie technologie en werkgelegenheid (pp. 21–32). Den Haag: Rathenau Instituut. Retrieved on 03.12.2015 from https://​www​.rathenau​.nl/​nl/​publicatie/​werken​-aan​-de​ -robotsamenleving​.Levy, F., and R. Murnane (2013). Dancing with Robots: Human Skills

22  Research handbook on human resource management and disruptive technologies

for Computerized Work. Washington, DC: Third Way NEXT. Retrieved on 03.12.2015 from http://​content​.thridwayorg/​publications/​715/​Dancing​-with​-Robots​.pdf. Manyika, J., M. Chui, P. Bisson, J. Woetzel, R. Dobbs, J. Bughin, and D. Aharon (2015). Unlocking the Potential of the Internet of Things. McKinsey Global Institute, 1 (2015). McKinsey Global Institute (2018). The Promised Challenges of the Age of Artificial Intelligence. McKinsey & Company. Miller, B., and R.D. Atkinson (2013). Are robots taking our jobs, or making them? Washington DC: The Information Technology & Innovation Foundation (ITIF). Retrieved on 24.04.2019 from https://​itif​.org/​publications/​2013/​09/​09/​are​-robots​-taking​-our​-jobs​-or​ -making​-them. Multu, B. (2011). Designing Embodied Cues for Dialogue with Robots. AI Magazine, 32 (4), 17–30. Pajarinen, M., and P. Rouvinen (2014). Computerization threatens one third of Finnish employment. Etla Brief, 22(13.1), 2014. Pajarinen, M., P. Rouvinen, and A. Ekeland (2015). Computerization and the Future of Jobs in Norway. Statistisk sentralbyrå. Retrieved on 31.01.2016 from http://​nettsteder. regjeringen. no/fremtidensskole/files/2014/05/Computerization-andthe-Future-of-Jobs-in-Norway.pdf. RAET (2016). RAET Benchmark Research into HR Trends. RAET. Retrieved on 21.07.2021 from https://​docplayer​.nl/​12096415​-Samenvatting​-raet​-hr​-benchmark​-2016​.html. Schattorie, J., A. De Jong, M. Fransen and B. Vennemann (2014). De impact van automatisering op de Nederlandse Arbeidsmarkt. Amstelveen: Deloitte. Schouteten, R. (2015). Robotisering: het kan, maar moet het ook? Tijdschrift voor Arbeidsvraagstukken, (31) 2, 124–146. [Robotization: it is possible, but should it be? Journal for Labor Issues.] Smith, A., and J. Anderson (2014). AI, Robotics and the Future of Jobs. Pew Research Center. Retrieved on 10.11.2020 from http://​www​.pewinternet​.org/​2014/​08/​06/​future​-of​-jobs/​. Van Est, R., and L. Kool (Eds.) (2015). Werken aan robotsamenleving: visies en inzichten uit de wetenschap over de relatie technologie en werkgelegenheid. Den Haag: Rathenau Instituut. Van der Zee, F. (2015). Technologie en arbeidsproductiviteit. In R. Van Est and L. Kool (Eds.), Werken aan de robotsamenleving: Visies en inzichten uit de wetenschap over de relatie technologie en werkgelegenheid (pp. 93–110). Den Haag: Rathenau Instituut. Retrieved on 03.12.2015 from https://​www​.rathenau​.nl/​nl/​publicatie/​werken​-aan​-de​-robotsamenleving. World Economic Forum (2016). The Future of Jobs. Employment, Skills and Workforce Strategy for the Fourth Industrial Revolution. January 2016. Ref 010116.

3. A self-determination theory framework to develop motivation-enhancing algorithmic management Xavier Parent-Rocheleau, Marylène Gagné and Antoine Bujold

INTRODUCTION The present chapter is about algorithmic management (AM) and its relationship with workers’ motivation. Algorithmic management is considered as one of the most disruptive HR-related technologies because it shifts the paradigm of the power dynamic between workers and technology from ‘technology as a tool’ to ‘technology as a boss’ (Gal et al., 2020; Kellogg et al., 2020; Parent-Rocheleau and Parker, 2022). Motivation is a key component of workers’ wellbeing and performance (Deci et al., 2017). For this reason, scholars and practitioners have devoted a great deal of attention to human resource management (HRM) practices capable of creating and sustaining high quality motivation among employees. However, important technological advances such as AM are changing the nature of work, jobs and tasks (Gagné, Parker et al., 2022; Hoffman et al., 2020). To preserve and enhance the optimal functioning and wellbeing of individuals, scholars have started to examine the repercussions of these rapid changes on the capacity of this new work design to maintain workers’ motivation (Gagné, Parker, et al., 2022). Algorithmic management (AM) has been defined in several ways (Duggan et al., 2019; Jabagi et al., 2019; Lee et al., 2015; Meijerink and Bondarouk, 2021; Möhlmann and Zalmanson, 2017). Despite their subtleties and differences, these definitions mostly refer to AM as a system of control where algorithms are responsible for making and executing decisions affecting workers, relying on intensive data collection of employees’ activities. Although still nascent, the literature on AM is developing at an impressive pace, with several theoretical propositions supported by a limited but growing number of empirical studies (Gagné, Parent-Rocheleau et al., 2022; Parent-Rocheleau and Parker, 2022). In these, the authors have mapped the main functions of AM, attempting to provide an empirically based answer to the question: What exactly is the role of algorithms in management? Kellogg et al. (2020) first proposed three categories of AM functions based on the typology of control mechanisms developed by Edwards (1979): algorithmic direction (recommending specific decisions or courses of action, restricting behaviours and access to information), algorithmic evaluation (recording information and rating workers) and algorithmic discipline (rewarding and replacing workers). Following this seminal work, 23

24  Research handbook on human resource management and disruptive technologies

other authors have proposed related but more granular or specific AM functions. For instance, Parent-Rocheleau and Parker (2022) built a conceptual model linking six AM functions (i.e. monitoring, scheduling, goal setting, performance management, compensation and job termination) to work design and job characteristics. Meijerink and Bondarouk (2021) also propose a set of six functions of algorithmic management of HR: recruitment, selection, training, appraisal, compensation and benefits, and workforce planning. These classifications suggest that the concept of AM presents unclear ties to closely related concepts, namely algorithmic HRM and algorithmic leadership. The term algorithmic HRM has been used to describe the increasing use of algorithms in HRM activities (Meijerink et al., 2021). Algorithmic leadership characterises the leader as either a machine or a person, focusing on the components of leadership roles and functions that can be automated through algorithmic systems (Wesche and Sonderegger, 2019). Besides describing algorithmic affordances in leaders’ roles, this literature also includes propositions about the role of human managers in a digitalised world, covering topics like the future of leadership skills, how to build people’s trust in technology and leadership in a human-machine teaming context (de Cremer, 2020). As illustrated in Figure 3.1, we argue that the concept of AM is at the frontier of algorithmic HRM and algorithmic leadership, since several of their respective topics of interest overlap, while also showing differences with both.

Figure 3.1

Differences and overlaps between AM and related constructs

This chapter will focus on the repercussions of AM as first described by Kellogg et al. (2020) (algorithmic direction, evaluation and discipline) on workers’ motivation. In a recent review of the empirical AM literature (Gagné, Parent-Rocheleau, et al., 2022), we observed that the use and characteristics of AM systems are likely to alter work motivation. We found empirical results highlighting negative worker

Framework to develop motivation-enhancing algorithmic management  25

experiences, such as lack of autonomy, distrust, injustice, stressful competition with coworkers, power asymmetry, dehumanisation of work, frustration, and anxiety, which may lead to a degradation of workers’ optimal functioning. The overall current evidence on AM suggests that its presence in organisations is associated with more negative than positive outcomes. However, the literature identifies promising solutions likely to mitigate these negative consequences and preserve or even enhance high quality jobs even under algorithmic management. In line with this constructive perspective, we contend that AM does not necessarily create negative motivational consequences regardless of the characteristics of the system and how organisations choose to use it to manage employees. The objective of the chapter is to propose a model to guide future studies on AM, future AM systems design and future AM organisational implementation. Our ultimate goal in doing so is to contribute to the deployment of more responsible AM that facilitates the conditions for workers’ optimal functioning through the satisfaction of basic human psychological needs. We use self-determination theory to inform this model and outline its main propositions in the next section.

SELF-DETERMINATION THEORY Self-determination theory (SDT) is a meta-theory of human motivation comprising several mini-theories that define motivation and the individual and environmental sources of influence on motivation (Deci and Ryan, 1985). The theory is supported by over 50 years of empirical research across different life domains, including the work domain (Deci et al., 2017). SDT proposes that motivation can take multiple forms, including intrinsic motivation (i.e. doing something for its own sake, out of interest and enjoyment) and extrinsic motivation (i.e., doing something for an instrumental reason). Extrinsic motivation can vary in terms of how internalised it is, while intrinsic motivation is completely internalised. Internalisation describes the process by which an external demand is adopted as one’s own and regulates a person’s behaviour (Deci et al., 2017). There are several forms of extrinsic motivation, including external regulation (doing something to obtain a reward or avoid a punishment controlled by others), introjected regulation (doing something to increase or maintain one’s self-worth or to avoid shame) and identified regulation (doing something because one finds it personally important or meaningful). These different forms of motivation are not equally related to workers’ optimal functioning. The more internalised forms of motivation, namely intrinsic and identified extrinsic motivation, are most strongly and positively associated with positive outcomes (e.g. work engagement, job satisfaction, commitment to the organisation, performance, proactivity) and also most strongly negatively associated with negative outcomes (e.g. distress, burnout, turnover intentions). Compared to these forms of motivation, accounting for over 60% of the variance in work-related outcomes, only 20% come from less internalised forms of motivation, i.e. external and introjected extrinsic motivation (Van den Broeck et al., 2021).

26  Research handbook on human resource management and disruptive technologies

SDT also proposes that human beings have three psychological needs that significantly influence their motivation (Deci and Ryan, 2000). The need for competence is defined as the need to feel effective when interacting with the environment (e.g. work tasks). The need for autonomy is defined as the need to feel volitional when behaving. The need for relatedness is defined as the need to feel connected to other people, to belong and to be cared for. All three needs have been shown to be independently crucial for the development and maintenance of more internalised motivation (i.e. identified and intrinsic motivation), and significantly related to important work outcomes, including performance, positive affect, work engagement, job satisfaction and organisational commitment (Van den Broeck et al., 2016). Consequently, research has concentrated on organisational practices that influence the satisfaction of these three psychological needs, finding that leadership and work design, among other things, are important considerations to ensure optimal employee motivation and work outcomes (Van den Broeck et al., 2016). We can use this rich SDT literature to inform how AM can be enhanced, namely how it can be designed and used in a way that promotes the satisfaction of these three psychological needs.

ALGORITHMIC MANAGEMENT AND SDT Previous research discusses or reveals the detrimental effect of the presence and magnitude of AM on workers’ autonomy (Galière, 2020; Jabagi et al., 2019; Leclercq-Vandelannoitte, 2017; Lee et al., 2015; Toyoda et al., 2020). In a recent review article, Gagné, Parker, et al. (2022) presented assumptions about the mechanisms through which current forms of AM yield more frustration than satisfaction of workers’ needs for autonomy, competence and relatedness. We elaborate on these assumptions below, before turning our attention to the core of our proposed model in the following section. Autonomy Needs The potential loss of autonomy for workers in the context of AM is one of the most documented topics in the field (see Parent-Rocheleau and Parker (2022) or Meijerink and Bondarouk (2021) for reviews). Found first among gig workers, this restriction of autonomy was initially a surprising finding considering the high level of freedom and flexibility commonly associated with freelancing and platform work. But researchers soon shed light on the ubiquitous algorithmic surveillance and control of task assignments, pay and performance ratings exerted by platforms that seriously limit this promised autonomy (Rosenblat, 2018). For instance, Möhlmann et al. (2021) observed tension between the autonomy provided by a flexible job as an Uber driver and the pervasive control deployed through algorithms. Regardless of the type of workplace where AM is implemented (platform or traditional economy), it relies on data. Data on various aspects of workers and productivity is collected through monitoring, surveillance, sensors, wearables and GPS,

Framework to develop motivation-enhancing algorithmic management  27

but also customer ratings, HR records, etc. Combined with other indicators (e.g. road traffic, customer traffic, weather forecast), this data is the fuel of AM systems, the key information used to make predictions and decisions, create schedules and nudges, set further productivity targets and promote, downgrade or terminate jobs. This extensive but necessary exposure to data-collection devices has been referred to as datafication (Schafheitle et al., 2020) or behavioural visibility (Leonardi and Treem, 2020). As a result of this data-based decision-making, workers try to ‘produce’ data that will lead to positive outcomes or favourable decisions, a phenomenon that we refer to as ‘working for data’. Focusing on tasks or behaviours that are monitored and quantified, workers may be taken away from the more meaningful or interesting parts of their jobs; for example, the nudges sent to Uber drivers to encourage them to log on at specific times, with the ratio of compliance to the nudge being one the indicators to establish performance scores (Rosenblat, 2018; Shalini and Bathini, 2021). This algorithmically induced orientation of one’s work is one of the ways through which AM may hamper the fulfilment of autonomy needs. Relatedness Needs The current forms of AM seem to be associated with more individualised work (Gagné, Parker et al., 2022). Two main reasons underlie this assumption. First, AM comes with less (in the traditional economy) or no (in the gig economy) contact with a human supervisor (Jarrahi et al., 2021; Walker et al., 2021). Second, when it comes to social contact with coworkers, studies have shed light on situations where AM can trigger a stressful competition climate between coworkers by displaying individual rankings (Leclercq-Vandelannoitte, 2017; Levy, 2015). This form of performance management may prevent gig workers from forging relationships with each other or erode existing relationships. For care workers, increasingly subject to AM (Bailey et al., 2020; Macdonald, 2021), the working-for-data phenomenon can limit meaningful and important social contact with patients. Moore and Hayes (2017) showed how electronic surveillance of homecare nurses increased productivity pressure (i.e. more patients to see) and led to a drastic reduction of time dedicated to psychosocial interactions. Competence Needs Early research and propositions provide mixed predictions about the potential repercussions of AM on the fulfilment of competence needs (Gagné, Parker et al., 2022). AM-induced performance indicators are precise, synchronous, clear and often directly linked to rewards (Möhlmann et al., 2021; Rosenblat, 2018). However, the precision, transparency and explainability of these indicators have been widely questioned (Duggan et al., 2019; Goods et al., 2019; Rani and Furrer, 2020). This lack of understandability of algorithmic performance feedback combined with frequent and unexpected updates or changes in criteria (Griesbach et al., 2019; Lee et al., 2015;

28  Research handbook on human resource management and disruptive technologies

Mäntymäki et al., 2019) are likely to seriously limit its contribution to the satisfaction of competence needs. In sum, early findings lead to several assumptions regarding the effect of AM on the satisfaction of psychological needs and motivation. Our aim is to provide constructive suggestions regarding key characteristics of algorithmic systems’ architecture and policies surrounding the use of a system capable of leveraging better satisfaction of psychological needs. Inspired by previous work advancing such ‘solutions’ (e.g. Jabagi et al., 2020; Parent-Rocheleau and Parker, 2022), we propose a model of motivation-enhancing AM (Figure 3.2) that specifies system characteristics as well as policies and practices likely to enhance the satisfaction of psychological needs and consequently promote self-determined motivation. The following sections elaborate on these key characteristics.

MOTIVATION-ENHANCING AM: CHARACTERISTICS OF THE SYSTEM Ensuring Transparency The transparency of algorithmic systems is receiving a great deal of attention from scholars. Transparency is considered key to the responsible use of artificial intelligence and as a solution to the opacity of machine learning-based decision-making, referred to as the black box problem (Langer and König, 2021; Zednik, 2021). But transparency of AM is also a complex and multifaceted construct. The most basic form of transparency is awareness of the presence of an algorithmic system in one’s work. Other important aspects proposed by authors are explainability or causability of decisions, understandability and data visibility or accessibility (Arrieta et al., 2020; Langer et al., 2021; Shin, 2021). All these aspects of transparency aim to leverage workers’ ability to understand how systems work (e.g. what is monitored) and how decisions are made. Opaque systems that do not allow workers such explanations are likely to deprive them of a sense of mastery in their work. For instance, it is common for Uber drivers to receive nasty surprises when they see their pay because algorithms lead to unpredictable compensation with very limited explanation and breakdown of the amount (Bokányi and Hannák, 2020; Möhlmann et al., 2021; Rosenblat, 2018). Such opacity generates uncertainty, an important burden on employees’ ability to define action in advance and control the course of their work (Gagné, Parker, et al., 2022). System transparency reduces uncertainty and allows employees to make sense of it to autonomously adapt to their environment. In SDT, offering a rationale about the why (not only the how) is one of the major interventions to increase feelings of autonomy (Steingut et al., 2017).

Framework to develop motivation-enhancing algorithmic management  29

Figure 3.2

Motivation-enhancing AM model

Keeping Humans in the Loop Most AM systems do not need human input other than data to accomplish their functions. However, being kept out of the loop of important decisions related to their work is likely to frustrate employees’ psychological needs. This could be caused by a feeling of dehumanisation of work (Crawford, 2021; Delfanti, 2021b; Guendelsberger, 2019) or a loss of dignity due to being managed by machines (Lamers et al., 2022). In a recent conceptual piece, Lamers et al. (2022) discuss the notion of workers’ capabilities in AM settings (freedom to learn, have social interaction, work efficiently and have financial stability), leveraging their ability to make decisions in their work. Giving workers power and a voice in the system seems essential to meet competence and autonomy needs. For instance, studies show that individuals are more likely to accept algorithmic decisions or recommendations if they are given the ability to make some decisions over the system (Dietvorst et al., 2016; Holland et al., 2017; Newman et al., 2020). The study by Newman et al. (2020) indicates that such power over automated decisions alleviates perceptions of reductionism associated with being managed by algorithms. Workers’ voices – or the ability to question or contest algorithmic decisions – may also promote workers’ agency. For instance, workers are often powerless when faced with unfair or erroneous poor ratings from customers, affecting their pay and future assignments (Rani and Furrer, 2020). An increasing number of studies show that workers develop workarounds or resistance strategies to counteract unfair or unintelligible decisions (Bakewell et al., 2018; Gregory, 2020; Heiland, 2021; Möhlmann and Zalmanson, 2017). Giving a voice to workers as part of the system design is likely to contribute to satisfying their psychological needs for competence and autonomy. The competence needs of human managers should also not be overlooked. Vargas (2021) provided a compelling illustration of how frontline managers are powerless in the face of algorithmic staffing and scheduling decisions. Frequent

30  Research handbook on human resource management and disruptive technologies

understaffing resulting from automated processes significantly limits their capacity to manage stores and handle customers’ complaints. Giving managers some say or power in staffing decisions would certainly result in a better sense of autonomy and competence. The study by Uhde et al. (2020) shows how such power over the system can be beneficial to workers whose schedules are assigned by an app when it allows them to collectively manage shift exchanges and self-manage the schedule’s creation through the app to fit their needs. Such flexibility and openness to collaboration was found to foster a team climate, which is likely to fulfil relatedness needs.

MOTIVATION-ENHANCING AM: POLICIES AND PRACTICES Limiting Surveillance and Monitoring The collection of data about work activities is at the core of AM. Monitoring (or surveillance) of workers is thus the essential entry point of most data needed by algorithms to make decisions, predictions or recommendations about schedule, pay, assignments, etc. (Kellogg et al., 2020). For example, activity-tracking software is becoming increasingly sophisticated, now even able to predict turnover by analysing emojis (Lu et al., 2022). There is also a growing variety of sensors, wearables or chips able to capture increasingly detailed and granular things (Schafheitle et al., 2020). Charbonneau and Doberstein (2020) show that remote workers’ acceptance of surveillance devices is negatively associated with their level of intrusiveness. In sum, the spread of AM occurs in the context of a shift from traditional forms of productivity monitoring (Ravid et al., 2020) to more pervasive forms of surveillance. The ubiquity of these new forms of surveillance raises important questions about its necessity or value. We argue that restricting monitoring and surveillance to direct indicators of productivity and necessary information would help to preserve intrinsic motivation by preventing the frustration of autonomy and competence needs. Authors have described how ubiquitous systems make workers feel controlled (Leclercq-Vandelannoitte, 2017; Leonardi and Treem, 2020; Schafheitle et al., 2020; Zorina et al., 2021). Monitoring and tracking only what is necessary is likely to reduce the working-for-data phenomenon described above, where workers lose their autonomy to choose to focus on the aspects of their jobs they value most. In other words, organisations must be careful not to fall into the trap of over monitoring, and AM should not be an excuse to bring Big Brother into the workplace. Managing Productivity Fairly Beyond the collection of data, fairness and reasonableness in the use of the data is also likely to exert a great influence on motivation. Monitoring data is often used to rate employees’ performance or to establish future performance targets. We propose

Framework to develop motivation-enhancing algorithmic management  31

three examples of how this process could be made fairer and more likely to satisfy rather than frustrate workers’ psychological needs. The first consists of limiting the dependency of workers’ performance ratings on customer evaluations. Call centres, hotels, restaurants, delivery companies, airlines and many others increasingly survey customers about their satisfaction regarding the service provided. This is nothing new, but AI-driven algorithms allow a more precise, systematic, timely and generalised use of customer satisfaction data (Bakewell et al., 2018; Evans and Kitchin, 2018; Galière, 2020; Gerber and Krzywdzinski, 2019; Meijerink, 2021). Due to the strong impact of negative evaluations on their pay, rankings, future assignments and job permanence ratings, workers have reported reaching a point where they would do anything to ensure good reviews, even if physically or mentally unsafe (Gregory, 2020; Griesbach et al., 2019). Considering the volatility, potential bias and questionable validity of customer ratings (Gerber and Krzywdzinski, 2019; Orlikowski and Scott, 2013; Rosenblat et al., 2017) and the lack of control workers may have over conditions that affect customer ratings (e.g. the weather), overreliance on these data contributes to workers’ powerlessness over their performance scores and thus on increasing uncertainty about their job and income, frustrating their needs for autonomy and competence. The second example consists of continually increasing performance targets based only on the performance of top performers. This situation, mostly reported in algorithmically managed warehouses, concerns the number of packages to be packed or handled in a delimited period of time (Delfanti, 2021a, 2021b; Guendelsberger, 2019). Productivity is tracked and optimised through workers’ barcode scanners, often connected to a wristband through which their movements are tracked (Delfanti, 2021a; Delfanti and Frey, 2021). Productivity targets are then based on the optimal sequence of movements leading to the highest number of items handled, which increases work intensity and decreases workers’ ability to regulate or organise their workload (Guendelsberger, 2019). In other words, workers are penalised if they don’t reach constantly increasing targets, but performing well also comes at a price. The third example is of systems that use metrics to compare and rank workers and deliberately induce stressful competition between them. Competition has been shown to reduce intrinsic motivation (Vallerand et al., 1986). In the gig economy, such competition is often used to control the supply of workers in a given territory (Gerber and Krzywdzinski, 2019; Lehdonvirta, 2018). However, an unintended consequence is that it deprives them of the often-marketed promise of freedom to choose where and when to work. In traditional work settings, performance metrics are often used to create within-team or company employee rankings (Evans and Kitchin, 2018; Leclercq-Vandelannoitte, 2017), but often result in triggering mental health problems, conflict between workers and employee turnover, especially when they are linked to financial incentives (Dahl and Pierce, 2020; Gläser et al., 2022) These consequences have been attributed to a decrease in the satisfaction of the three psychological needs (Gagné, Parker, et al., 2022; Kuvaas et al., 2020).

32  Research handbook on human resource management and disruptive technologies

Maintaining Pay Stability Algorithmic management has widely been associated with higher uncertainty and instability in workers’ earnings (Goods et al., 2019; Griesbach et al., 2019; Jarrahi et al., 2021; Peticca-Harris et al., 2020; Rani and Furrer, 2020; Vargas, 2021). This is mainly because AM tends to facilitate pay-for-performance (PfP). Authors have described the different direct or indirect forms of PfP in AM, such as piecework pay, tipping policies, algorithmic scheduling and work assignment-based pay. First, AM makes it easy to deconstruct work into discrete tasks and remunerate each task achieved, which has commonalities with the historic piecework form of compensation (Alkhatib et al., 2017), where paid time is solely limited to productive time, consequently excluding travel or wait time (Conroy et al., 2021; Degryse, 2020). Second, in the gig economy, tips are a fundamental component of workers’ income and are directly contingent on customer satisfaction (Duhaime and Woessner, 2019); yet, platforms exert different levels of control and transparency over the tips offered by customers, such that workers often do not know the tip amount related to a specific task (Myhill et al., 2021; Rosenblat, 2018). Third, algorithmically managed companies often reward good performers with more profitable schedules and better task assignment (Gerber and Krzywdzinski, 2019; Gregory, 2020; Griesbach et al., 2019), which not only exacerbates the impact of performance on earnings but also creates greater pay dispersion between workers. SDT has a long record of studies showing the negative effect of PfP on self-determined motivation (Gagné and Forest, 2008; Kuvaas et al., 2020). Performance-based compensation tends to render work more instrumental and encourage people to seek rewards rather than meaningfulness in their job, reducing their agency and sense of autonomy. Also, researchers found that PfP may encourage aggressive competition among coworkers and deviant behaviours (Gläser et al., 2017, 2022; Lee, 2020), which is likely to frustrate relatedness needs. In addition to PfP, studies have shed light on the issue of pay unpredictability associated with AM, mostly in the gig economy. This may be due to the frequent changes in pay calculation processes of algorithmic systems (Bokányi and Hannák, 2020; Lee et al., 2015; Mäntymäki et al., 2019) or to the dynamic pricing deployed by some platforms that modulate the price of services based on real-time demand, often without disclosing rates to workers (Seele et al., 2021; Shapiro, 2020). Given that pay can be an indicator of one’s performance (i.e. a form of feedback), such unpredictability can reduce feelings of competence. However, AM does not necessarily come with a high degree of PfP or pay unpredictability. These are practices and policies that organisations using AM decide to deploy or not. Keeping a low to moderate degree of PfP and ensuring predictability of pay in the context of AM can help create and sustain self-determined motivation among workers through the satisfaction of psychological needs.

Framework to develop motivation-enhancing algorithmic management  33

Providing Constructive Feedback Along with performance metrics, algorithmic systems can provide detailed, synchronous, clear and frequent feedback to workers. This feedback is mostly automated and delivered electronically. Early evidence on algorithmic performance feedback has yielded mixed predictions (Parent-Rocheleau and Parker, 2022). Algorithmic feedback can promote autonomous motivation when bias-free, transparent, informative, developmental and not aimed to create competition (Gagné, Parker, et al., 2022). For instance, Meijerink (2021) developed a multilevel model of online consumer reviews detailing the conditions and mechanisms through which consumer reviews can constitute valuable feedback for gig workers, helping them to craft their work. That said, in traditional work settings, the extent to which AM manages to replace the humanness of a manager’s feedback remains an important question. Preserving Relationships Algorithmic management can exert an individualising effect on work, namely due to PfP and individual rankings (Gagné, Parker, et al., 2022), but also because it reduces the need for contact with human managers, whose roles it changes drastically (Walker et al., 2021). For this reason, it is essential in AM settings to find ways to maintain and promote social relationships. This can be done through algorithmic systems, like in the study by Uhde et al. (2020) revealing how the scheduling app facilitated proactive collaboration between care workers. It can also be done by encouraging and supporting groups and communities of workers. This latter phenomenon is largely found in the gig economy, where large communities are created by workers and serve as important places for social interaction, networking and sensemaking (Heiland, 2021; Toth et al., 2020; Walker et al., 2021). We argue that organisations, whether they are platforms or traditional, can leverage self-determined motivation by promoting and supporting the creation of such communities.

CONCLUSION AND RECOMMENDATIONS FOR ACTION Our goal in this chapter was to contribute to a shift in the research on algorithmic management from enumerating its consequences for workers to providing concrete ways to guide a more human and responsible implementation. We suggested a model with solutions for more motivation-enhancing forms of AM. We believe that self-determined motivation is key to workers’ optimal functioning at work, encompassing mental health, wellbeing and fulfilment. If AM is to be part of the future of work, it will have to be motivation-enhancing.

34  Research handbook on human resource management and disruptive technologies

REFERENCES Alkhatib, A., M.S. Bernstein, and M. Levi (2017). Examining crowd work and gig work through the historical lens of piecework. In Proceedings of the 2017 CHI Conference on Human Factors in Computing Systems. Arrieta, A.B., N. Díaz-Rodríguez, J. Del Ser, A. Bennetot, S. Tabik, A. Barbado, S. García, S. Gil-López, D. Molina, and R. Benjamins (2020). Explainable Artificial Intelligence (XAI): Concepts, taxonomies, opportunities and challenges toward responsible AI. Information fusion, 58, 82–115. Bailey, S., D. Pierides, A. Brisley, C. Weisshaar, and T. Blakeman (2020). Dismembering organisation: The coordination of algorithmic work in healthcare. Current Sociology, 68(4), 546–71. Bakewell, L.L., K. Vasileiou, K.S. Long, M. Atkinson, H. Rice, M. Barreto, J. Barnett, M. Wilson, S. Lawson, and J. Vines (2018). Everything we do, everything we press: Data-driven remote performance management in a mobile workplace. In Proceedings of the 2018 CHI Conference on Human Factors in Computing Systems. Bokányi, E., and A. Hannák (2020). Understanding inequalities in ride-hailing services through simulations. Scientific Reports, 10(1), 1–11. Charbonneau, É., and C. Doberstein (2020). An empirical assessment of the intrusiveness and reasonableness of emerging work surveillance technologies in the public sector. Public Administration Review, 80(5), 780–91. Conroy, S.A., D. Roumpi, J.E. Delery and N. Gupta (2021). Pay volatility and employee turnover in the trucking industry. Journal of Management, Online first publication. Crawford, K. (2021). The Atlas of AI. Yale University Press. Dahl, M.S., and L. Pierce (2020). Pay-for-performance and employee mental health: Large sample evidence using employee prescription drug usage. Academy of Management Discoveries, 6(1), 12–38. de Cremer, D. (2020). Leadership by algorithm. Harriman House. Deci, E.L., A.H. Olafsen, and R.M. Ryan (2017). Self-determination theory in work organizations: The state of a science. Annual Review of Organizational Psychology and Organizational Behavior, 4, 19–43. Deci, E.L., and R.M. Ryan (1985). The general causality orientations scale: Self-determination in personality. Journal of Research in Personality, 19(2), 109–34. Deci, E.L., and R.M. Ryan (2000). The ‘what’ and ‘why’ of goal pursuits: Human needs and the self-determination of behavior. Psychological Inquiry, 11(4), 227–68. Degryse, C. (2020). Du flexible au liquide: le travail dans l’économie de plateforme. Relations industrielles/Industrial Relations, 75(4), 660–83. Delfanti, A. (2021a). Machinic dispossession and augmented despotism: Digital work in an Amazon warehouse. New Media & Society, 23(1), 39–55. Delfanti, A. (2021b). The Warehouse: Workers and Robots at Amazon. Pluto Press. Delfanti, A., and B. Frey (2021). Humanly extended automation or the future of work seen through Amazon patents. Science, Technology, & Human Values, 46(3), 655–82. Dietvorst, B.J., J.P. Simmons and C. Massey (2016). Overcoming algorithm aversion: People will use imperfect algorithms if they can (even slightly) modify them. Management Science, 64(3), 1155–70. Duggan, J., U. Sherman, R. Carbery and A. McDonnell (2019). Algorithmic management and app-work in the gig economy: A research agenda for employment relations and HRM. Human Resource Management Journal, Online Publication First. Duhaime, E.P., and Z.W. Woessner (2019). Explaining the decline of tipping norms in the gig economy. Journal of Managerial Psychology, 34(4), 233–45. Edwards, R. (1979). Contested terrain: The transformation of the workplace in the twentieth century. Basic Books.

Framework to develop motivation-enhancing algorithmic management  35

Evans, L., and R. Kitchin (2018). A smart place to work? Big data systems, labour, control and modern retail stores. New Technology, Work and Employment, 33(1), 44–57. Gagné, M., and J. Forest, (2008). The study of compensation systems through the lens of self-determination theory: Reconciling 35 years of debate. Canadian Psychology, 49(3), 225–32. Gagné, M., X. Parent-Rocheleau, A. Bujold, M.-C. Gaudet and P. Lirio (2022). How algorithmic management influences worker motivation: A self-determination theory perspective. Canadian Psychology, 63(2), 247–60. Gagné, M., S.K. Parker, M.A. Griffin, P.D. Dunlop, C. Knight, F. Klonek and X. Parent-Rocheleau (2022). Understanding and shaping the future of work with self-determination theory. Nature Reviews Psychology. https://​doi​.org/​https://​doi​.org/​10​.1038/​s44159–022–00056​-w. Gal, U., T.B. Jensen and M.-K. Stein (2020). Breaking the vicious cycle of algorithmic management: A virtue ethics approach to people analytics. Information and Organization, 30(2), 1–15. Galière, S. (2020). When food-delivery platform workers consent to algorithmic management: A Foucauldian perspective. New Technology, Work and Employment, Online publication First. Gerber, C., and M. Krzywdzinski (2019). Brave new digital work? New forms of performance control in crowdwork. In S.P. Vallas and A. Kovalainen (Eds.), Work and Labor in the Digital Age (pp. 121–43). Emerald Publishing Limited. Gläser, D., S. van Gils and N. Van Quaquebeke (2017). Pay-for-Performance and interpersonal deviance. Journal of Personnel Psychology, 16(2), 77–90. Gläser, D., S. van Gils, and N. Van Quaquebeke (2022). With or against others? Pay-for-Performance activates aggressive aspects of competitiveness. European Journal of Work and Organizational Psychology, 1–15. https://​doi​.org/​10​.1080/​1359432X​.2022​ .2039125 Goods, C., A. Veen, and T. Barratt (2019). ‘Is your gig any good?’ Analysing job quality in the Australian platform-based food-delivery sector. Journal of Industrial Relations, 61(4), 502–27. Gregory, K. (2020). ‘My life is more valuable than this’: Understanding risk among on-demand food couriers in Edinburgh. Work, Employment and Society, Online first publication. Griesbach, K., A. Reich, L. Elliott-Negri and R. Milkman (2019). Algorithmic control in platform food delivery work. Socius, 5, 1–15. Guendelsberger, E. (2019). On the Clock. Little, Brown and Company. Heiland, H. (2021). Controlling space, controlling labour? Contested space in food delivery gig work. New Technology, Work and Employment, Online first publication. Hoffman, B.J., M.K. Shoss, and L.A. Wegman, (2020). The Cambridge Handbook of the Changing Nature of Work. Cambridge University Press. Holland, C., J. Levis, R. Nuggehalli, B. Santilli and J. Winters (2017). UPS optimizes delivery routes. Interfaces, 47(1), 8–23. Jabagi, N., A.-M. Croteau, L.K. Audebrand and J. Marsan (2019). Gig-workers’ motivation: Thinking beyond carrots and sticks. Journal of Managerial Psychology, 34(4), 192–213. Jabagi, N., A.-M. Croteau and L.K. Audebrand (2020). Perceived organizational support in the face of algorithmic management: A conceptual model. In Proceedings of the 53rd Hawaii International Conference on System Sciences. Jarrahi, M.H., G. Newlands, M.K. Lee, C.T. Wolf, E. Kinder and W. Sutherland (2021). Algorithmic management in a work context. Big Data & Society, 8(2), 1–14. Kellogg, K., M. Valentine and A. Christin (2020). Algorithms at work: The new contested terrain of control. Academy of Management Annals, 14(1), 366–410. Kuvaas, B., R. Buch, and A. Dysvik (2020). Individual variable pay for performance, controlling effects, and intrinsic motivation. Motivation and Emotion, 44(4), 525–33.

36  Research handbook on human resource management and disruptive technologies

Lamers, L., J. Meijerink, G. Jansen and M. Boon (2022). A capability approach to worker dignity under algorithmic management. Ethics and Information Technology, 24(1), 1–15. Langer, M., and C.J. König, (2021). Introducing a multi-stakeholder perspective on opacity, transparency and strategies to reduce opacity in algorithm-based human resource management. Human Resource Management Review, Online First Publication. Langer, M., D. Oster, T. Speith, H. Hermanns, L. Kästner, E. Schmidt, A. Sesing and K. Baum (2021). What do we want from Explainable Artificial Intelligence (XAI)? – A stakeholder perspective on XAI and a conceptual model guiding interdisciplinary XAI research. Artificial Intelligence, 296, 1–24. Leclercq-Vandelannoitte, A. (2017). An ethical perspective on emerging forms of ubiquitous IT-based control. Journal of Business Ethics, 142(1), 139–54. Lee, H.-W. (2020). Interpersonal competition in organization: An investigation of antecedents. International Journal of Manpower, 41(8), 1363–83. Lee, M.K., D. Kusbit, E. Metsky and L. Dabbish (2015). Working with machines: The impact of algorithmic and data-driven management on human workers. In Proceedings of the 33rd Annual ACM Conference on Human Factors in Computing Systems. Lehdonvirta, V. (2018). Flexibility in the gig economy: Managing time on three online piecework platforms. New Technology, Work and Employment, 33(1), 13–29. Leonardi, P.M., and J.W. Treem, (2020). Behavioral visibility: A new paradigm for organization studies in the age of digitization, digitalization, and datafication. Organization Studies, 41(12), 1601–25. Levy, K.E. (2015). The contexts of control: Information, power, and truck-driving work. The Information Society, 31(2), 160–74. Lu, X., W. Ai, Z. Chen, Y. Cao and Q. Mei (2022). Emojis predict dropouts of remote workers: An empirical study of emoji usage on GitHub. PLOS ONE, 17(1). https://​doi​.org/​10​.1371/​ journal​.pone​.0261262. Macdonald, F. (2021). ‘Personalised risk’ in paid care work and the impacts of ‘gig economy’ care platforms and other market-based organisations. International Journal of Care and Caring, 5(1), 9–25. Mäntymäki, M., A.., Baiyere and A.N. Islam (2019). Digital platforms and the changing nature of physical work: Insights from ride-hailing. International Journal of Information Management, 49, 452–60. Meijerink, J. (2021). Talent management in the gig economy: A multilevel framework highlighting how customers and online reviews are key for talent identification. In S. Wiblen (Ed.), Digitalised Talent Management: Navigating the Human-Technology Interface (pp. 98–121). Routledge. Meijerink, J., and T. Bondarouk (2021). The duality of algorithmic management: Toward a research agenda on HRM algorithms, autonomy and value creation. Human Resource Management Review, Online First Publication. Meijerink, J., M. Boons, A. Keegan and J. Marler (2021). Algorithmic human resource management: Synthesizing developments and cross-disciplinary insights on digital HRM. The International Journal of Human Resource Management, 1–18. Möhlmann, M., and L. Zalmanson (2017, December 10–13). Hands on the wheel: Navigating algorithmic management and uber drivers’ autonomy. In Proceedings of the International Conference on Information Systems (ICIS 2017), Seoul, South Korea. Möhlmann, M., L. Zalmanson, O. Henfridsson., and R.W. Gregory (2021). Algorithmic management of work on online labor platforms: When matching meets control. MIS Quarterly, 45(4), 1999–2022. Moore, S., and L. Hayes (2017). Taking worker productivity to a new level? Electronic Monitoring in homecare – the (re)production of unpaid labour. New Technology, Work and Employment, 32(2), 101–14.

Framework to develop motivation-enhancing algorithmic management  37

Myhill, K., J. Richards and K. Sang (2021). Job quality, fair work and gig work: The lived experience of gig workers. The International Journal of Human Resource Management, 32(19), 4110–35. Newman, D.T., N.J. Fast and D.J. Harmon (2020). When eliminating bias isn’t fair: Algorithmic reductionism and procedural justice in human resource decisions. Organizational Behavior and Human Decision Processes, 160, 149–67. Orlikowski, W.J., and S.V. Scott (2013). What happens when evaluation goes online? Exploring apparatuses of valuation in the travel sector. Organization Science, 25(3), 868–91. Parent-Rocheleau, X., and S.K. Parker (2022). Algorithms as work designers: How algorithmic management influences the design of jobs. Human Resource Management Review, 32(3), 1000838. Peticca-Harris, A., N. DeGama and M. Ravishanka, (2020). Postcapitalist precarious work and those in the ‘drivers’ seat: Exploring the motivations and lived experiences of Uber drivers in Canada. Organization, 27(1), 36–59. Rani, U., and M. Furrer (2020). Digital labour platforms and new forms of flexible work in developing countries: Algorithmic management of work and workers. Competition & Change, Online Publication First. Ravid, D.M., D.L. Tomczak, J.C. White and T.S. Behrend (2020). EPM 20/20: A review, framework, and research agenda for electronic performance monitoring. Journal of Management, 46(1), 100–26. Rosenblat, A. (2018). Uberland: how algorithms are rewriting the rules of work. University of California Press. Rosenblat, A., K.E. Levy, S. Barocas and T. Hwang (2017). Discriminating tastes: Uber’s customer ratings as vehicles for workplace discrimination. Policy & Internet, 9(3), 256–79. Schafheitle, S.D., A. Weibel, I.L. Ebert, G. Kasper, C. Schank and U. Leicht-Deobald (2020). No stone left unturned? Towards a framework for the impact of datafication technologies on organizational control. Academy of Management Discoveries, In-Press. Seele, P., C. Dierksmeier, R. Hofstetter and M.D. Schultz (2021). Mapping the ethicality of algorithmic pricing: A review of dynamic and personalized pricing. Journal of Business Ethics, 170(4), 697–719. Shalini, P., and D.R. Bathini (2021). Microtargeting control: Explicating algorithmic control and nudges in platform-mediated cab driving in India. New Technology, Work and Employment, 36(1), 74–93. Shapiro, A. (2020). Dynamic exploits: Calculative asymmetries in the on-demand economy. New Technology, Work and Employment, 35(2), 162–77. Shin, D. (2021). The effects of explainability and causability on perception, trust, and acceptance: Implications for explainable AI. International Journal of Human-Computer Studies, 146, 1–10. Steingut, R.R., E.A. Patall and S.S. Trimble (2017). The effect of rationale provision on motivation and performance outcomes: A meta-analysis. Motivation Science, 3(1), 19–50. Toth, I., S. Heinänen and K. Blomqvist (2020). Freelancing on digital work platforms – Roles of virtual community trust and work engagement on person–job fit. VINE Journal of Information and Knowledge Management Systems, 50(4), 553–67. Toyoda, Y., G. Lucas and J. Gratch (2020). The effects of autonomy and task meaning in algorithmic management of crowdwork. In Proceedings of the 19th International Conference on Autonomous Agents and MultiAgent Systems. Uhde, A., N. Schlicker, D.P. Wallach and M. Hassenzahl (2020). Fairness and decision-making in collaborative shift scheduling systems. In Proceedings of the 2020 CHI Conference on Human Factors in Computing Systems. Vallerand, R.J., L.I. Gauvin, and W.R. Halliwell (1986). Negative effects of competition on children’s intrinsic motivation. The Journal of Social Psychology, 126(5), 649–56.

38  Research handbook on human resource management and disruptive technologies

Van den Broeck, A., D.L. Ferris, C.-H. Chang and C.C. Rosen (2016). A review of self-determination theory’s basic psychological needs at work. Journal of Management, 42(5), 1195–229. Van den Broeck, A., J.L. Howard, Y. Van Vaerenbergh, H. Leroy and M. Gagné (2021). Beyond intrinsic and extrinsic motivation: A meta-analysis on self-determination theory’s multidimensional conceptualization of work motivation. Organizational Psychology Review, 11(3), 240–73. Vargas, T.L. (2021). Consumer redlining and the reproduction of inequality at dollar general. Qualitative Sociology, 44(2), 205–29. Walker, M., P. Fleming and M. Berti (2021). ‘You can’t pick up a phone and talk to someone’: How algorithms function as biopower in the gig economy. Organization, 28(1), 26–43. Wesche, J.S., and A. Sonderegger (2019). When computers take the lead: The automation of leadership. Computers in Human Behavior, 101, 197–209. Zednik, C. (2021). Solving the black box problem: a normative framework for explainable artificial intelligence. Philosophy & Technology, 34(2), 265–88. Zorina, A., F. Bélanger, N. Kumar and S. Clegg (2021). Watchers, watched, and watching in the digital age: Reconceptualization of information technology monitoring as complex action nets. Organization Science, 32(6), 1571–96.

4. Mitigating bias in AI-powered HRM Melika Soleimani, James Arrowsmith, Ali Intezari and David J. Pauleen

INTRODUCTION The use of artificial intelligence (AI) is rapidly increasing in human resource management (HRM) (Cheng and Hackett, 2021). First, AI applications are employed in recruitment and selection to analyse textual and visual data for purposes such as job description optimisation, targeted job advertising, multi-database candidate sourcing, CV screening, psychometric testing, video screening, background checking and candidate engagement via chatbot (Albert, 2019). For example, IBM Watson was developed as an AI and cognitive talent management solution to improve talent acquisition, matching job seekers’ skills and organisational needs and interacting with candidates (IBM, 2018). Second, in work management, AI is used in task ordering and labour scheduling, performance monitoring and rating, and employee discipline (Kellogg et al., 2020). Amazon is an exemplar company case but algorithmic work management systems are widely used across the logistics, manufacturing, retail, hospitality and call centre sectors. The shift to home-working under Covid also accelerated deployment across white-collar occupations (EU-OSHA, 2022). Third, AI is being increasingly used in the people analytics area to predict employee engagement, training, turnover intentions and use of incentives (Cheng and Hackett, 2021). This can be based on existing and past employee data. Proprietorial tools such as ‘SWOOP analytics’ can also be used to explore collaboration and behavioural insights by analysing the content and relationships in digital workplaces such as Microsoft Teams, Yammer and Workplace from Meta (Hüllmann et al., 2021). In each of these areas AI offers a breakthrough technology to improve HRM decisions by making sense of information more quickly and consistently than the capacity or inclination of humans (Deshpande et al., 2021). However, there are growing concerns about inadequate accountability, oversight and potential biases (Eurofound, 2022). Unfortunately, ‘while algorithms are often viewed as objective and impartial, they in fact have the potential to encode and amplify existing biases’ (World Economic Forum, 2021, p. 18). This is because the training data used for machine learning might not be of sufficient scale or representativeness if, for example, they are based on samples such as older white males. Measures used to appraise employees or job applicants can also be overly simplistic and the impact of wider social inequalities, including access to education, neglected in design (European Parliament, 2020). Recent research indicates that automation of the hiring process imposes narrow and 39

40  Research handbook on human resource management and disruptive technologies

inflexible criteria, to the mutual disadvantage of employers and applicants (Fuller et al., 2021). This can result in discrimination and a lack of diversity. According to Schwartz et al. (2022), potential bias in the development process of AI can be categorised into three main groups: systemic bias, statistical and computational bias and human bias. Systemic bias is the result of the historical, societal and institutional procedures and practices that value certain social groups and devalue minorities (Yapo and Weiss, 2018; Schwartz et al., 2022). Statistical and computational bias happens when datasets required for training machine learning (ML) models are not representative of the population (Ntoutsi et al., 2020). Data collection methods fail to obtain sufficient data for certain groups, especially those who have traditionally been marginalised (Smith and Rustagi, 2020). This bias also includes inappropriate and inaccurate statistical analysis caused by missing, inconsistent and incorrect values in datasets (Lones, 2021). Biased data might be a product of human decisions around model selection and metrics that emphasise accuracy and optimization over fairness (Forde et al., 2021). Hence the datasets used by ML processes, the decisions made by those who build ML models and those who provide the insight and oversight to make such systems actionable can be tied to human biases (Schwartz et al., 2022). Using AI in HRM requires the HR function to be aware of the risks associated with algorithmic bias and ML (Tambe et al., 2019). ML is the core part of AI that supports and enables software applications for computer vision, speech recognition, natural language processing and robotics (Jordan and Mitchell, 2015). It is also the application of mathematical models to perform tasks associated with prediction, classification, anomaly detection, segmentation and association (Strohmeier, 2022). Examples of ML models that support HR functions include those predicting future employee turnover (Strohmeier and Piazza, 2015), classifying employees into ordinal performance classes, e.g. ranging from ‘needs improvement’ to ‘outstanding’ (Kirimi and Moturi, 2016), and detecting employee performance anomalies (Lukashin et al., 2021). Each involves the risk of bias.

POTENTIAL BIASES IN DEVELOPING AI FOR HRM As noted above, there are three main sources of bias in developing AI. In the following sections, we further explain how these three types of bias occur in developing AI and its applications in HRM. Table 4.1 outlines the biases, with examples explored below. Systematic Bias Systematic bias is associated with datasets and assumptions used in algorithmic models that result in biases against racial and ethnic minorities (Angwin et al., 2016). Training datasets, in particular those scraped from internet sources, might have common gender, racial, cultural and socioeconomic biases (Bender et al., 2021). In

Mitigating bias in AI-powered HRM  41

HRM, systemic biases in datasets result in certain groups of applicants being unfairly disadvantaged. For example, algorithms for facial recognition have been shown to have difficulties identifying the expressions of black applicants, while automated voice analysis may have problems in understanding differences in speech patterns and vocabulary that correlate with race or ethnicity (Friedman and McCarthy, 2020). Datasets for performance reviews of high-performing employees in technology occupations (‘tech’) include biases against women who, while using the same language as men, are given negative personality ratings and deemed to be abrasive, strident, irrational or difficult to work with (Houser, 2019). In developing algorithmic models, improperly utilising protected attributes, including gender, age, race or religion, as proxies results in harmful biases (Kamiran and Calders, 2012; Malik, 2020). The term proxy defines a feature associated with memberships in a protected class that is used in decision-making and results in biased decisions (Datta et al., 2017). Proxies in algorithmic hiring are indirect variables used to measure employment suitability, including length of time in prior employment, productivity and the number of lost hours. Additionally, predictor variables including distance from employment location may be included as these may correlate with employee likelihood to quit their job due to long commutes or bad traffic. However, these proxies can contribute to AI biased outcomes. For example, ‘distance from the employment site’ disadvantages those on lower incomes who cannot move closer to work (Schwartz et al., 2022). An online tech hiring platform called Gild used ‘social data’ as well as other resources such as résumés to rank candidates in tech. ‘Social data’ is used as a proxy for sociability and collaboration that measures time spent developing and sharing codes on platforms such as GitHub and communicating with others online. However, women generally have less time to chat online due to childcare and family responsibilities. In fact, their acceptance rates and contributions on GitHub are higher when they do not declare as women (Terrell et al., 2017). In some AI models classification or association rules excluding information such as gender, age, or birth country about the candidate may not necessarily produce unbiased and fair outcomes (Pedreschi et al., 2008). For example, in one case a model that predicted the GPA of college students was shown to be efficient, yet biased against black students. When the researchers considered race, the algorithm admitted a more equitable percentage of white and black students with higher grades on average (Kleinberg et al., 2018). Statistical and Computational Bias Statistical and computational bias stems from errors rooted in the representation of complex data in mathematical models, heterogeneous data and issues with the treatment of outliers, data cleaning and imputation factors. Statistical and computational bias manifests in datasets as selection/exclusion bias and reporting bias. Selection/ exclusion bias happens when proper randomisation is not achieved, which results in the exclusion of appropriate and important individuals, groups or data from the

42  Research handbook on human resource management and disruptive technologies

analysis of the population of interest (Berk, 1983). A dataset is supposed to reflect the world accurately; however, there are significant gaps in datasets, including insufficient data from particular groups of people, which result in biased datasets (Crawford, 2013). For example, in an automated interview, algorithms cannot initially classify a candidate with a speech impediment since a sufficiently representative set of data does not exist (Lebovitz et al., 2021). Reporting bias can be considered as a type of selection bias when selection of certain kinds of results is likely to be over- or under-reported (Ntoutsi et al., 2020; Smith and Rustagi, 2020). For example, in the United States, federal and state authorities have difficulties capturing labour force data from Native American communities. The Native Americans’ data is often not standardised as different government databases use different criteria to identify tribal members and often misclassify race and ethnicity. This results in reporting bias in unemployment that undercounts the American Indian population (Smith and Rustagi, 2020). Statistical and computational bias appears in developing ML models while mathematically encoding the training data (Cowgill, 2018). AI developers often use a minimum number of parameters to make ML models simpler, more explainable, transparent and less expensive to build. However, simple models might worsen statistical biases due to limited assumptions on the training data that might not account for nuanced demographics (Schwartz et al., 2022). To overcome the issue of nonrepresentative datasets, AI developers might use data aggregation, in which raw data about multiple individuals are gathered and summarised as a form of statistical analysis to make predictions about individual behaviour. However, this approach can lead to biased outcomes (Angwin et al., 2016). This bias occurs as the result of making an inference about an individual based on their membership within a group, for example, predicting college students’ academic performance based on an individual’s race (Feathers, 2021). The risk of statistical and computational bias is compounded by overly focusing on selecting the most accurate ML models to optimise ML performance (Forde et al., 2021). For example, to improve the accuracy of ML models (i.e. classification models), one strategy is to increase the True Positive Rate while lowering error rates such as the False Positive Rate. True Positive is the number of positive instances correctly classified and False Positive is the number of negative instances incorrectly classified (Tchakounté and Hayata, 2017). In hiring algorithms, when the True Positive Rate for males is higher than for females, ML models select male profiles rather than relevant female profiles (Delecraz et al., 2022). In addition, AI-based hiring systems that claim to collect data about candidates from audio and video have been shown to increase bias in outcome decisions and may present untenable trade-offs between bias mitigation and prediction accuracy (Booth et al., 2021). Human Biases Human biases might arise from choices around how people collect and categorise data, including the determination of parameters for a dataset (Smith and Rustagi,

Mitigating bias in AI-powered HRM  43

2020). This also involves labelling and interpretation (Sun et al., 2020; Smith and Rustagi, 2020). For example, the German public radio outlet BR24 tested an AI system that creates a personality profile for use in hiring processes through assessing tone of voice, language, gestures and facial expressions (Harlen and Schnuck, 2021). The AI system scored candidates differently according to non-job-related factors such as wearing spectacles or having a bookshelf in the background, replicating biases from previous selection decisions. Another example of human bias in labelling datasets is when some words are labelled as positive words representing successful candidates. Amazons’ algorithms favoured applicants whose résumé included ‘strong’ words like ‘executed’ or ‘captured’, which were more commonly found in those from men (Manyika et al., 2019). Much of this bias transfers from management and HR decision-makers. Human bias can be implicit or unconscious and AI developers (who are overwhelmingly male) can also transfer their own biases into ML models when making decisions on what data points to include or exclude (Slota et al., 2021; Cowgill et al., 2020). Developers might be attentive to certain aspects of gender, race or sexual orientation and ensure that direct measures are excluded. However, this does not necessarily ensure that algorithms are neutral or nondiscriminatory. Mitigating bias in AI for HRM requires HR to assume a certain level of oversight and responsibility as it is not only a technical issue. Additionally, AI developers need to be rigorous in eliminating bias and in their pursuit of mitigation techniques including both technical and nontechnical means. In the following section we discuss some of the main techniques that can be used to mitigate biases in developing AI for HRM by focusing on HR and AI developers’ roles.

MITIGATING AI BIASES Applying appropriate technical solutions is not the only path to understanding bias and mitigation techniques (Charlwood and Guenole, 2022). Placing humans at the centre of the development process of AI through active participation in the preparation, learning and decision-making phases of AI assists in mitigating biases (Margetis et al., 2021). Thus, in the context of HRM, developing AI based on thorough HR involvement and partnership can decrease the risks of unfair and biased AI systems that negatively affect workers (Charlwood and Guenole, 2022). According to Moore (2020), HR can recommend protocols to guide the development of AI for HRM, including requirements for AI developers to have diverse engineering teams where possible (Charlwood and Guenole, 2022; Cowgill et al., 2020). Errors can also be pointed out, such as, for example, focusing on years of professional experience, which disadvantages women due to pregnancy and maternity-related leave (Delecraz et al., 2022). Certainly, too, AI developers need to develop standards for training datasets and AI models. These standards can assist in ensuring the availability of less biased training datasets and data curation methods (Jo and Gebru, 2020). Additionally, these standards clearly explain which features are important in producing AI outputs. This

44  Research handbook on human resource management and disruptive technologies

Table 4.1

Biases in developing AI for HRM

Sources of biases Type of biases

Reasons biases happen in AI

Example of biases in AI-HRM

● Socioeconomic biases in datasets

● Misidentifying facial expressions

in AI Datasets

Systemic bias

due to historical, societal and insti-

of minority applicants, analysing

tutional practices

differences in speech patterns due to race and ethnicity ● Biases against women in performance review due to ascribed personality factors

Statistical and computational bias

● Nonrepresentative datasets due to selection/exclusion bias and reporting bias

● Lack of data about ‘unusual’ candidates (e.g. speech impediment) ● Undercounting the Native American population in unemployment reports

Human bias

● Human biases in collecting and

● Scoring candidates who wore

labelling datasets are embedded

spectacles or had a bookshelf in

in AI

the background higher than other candidates ● Labelling some words such as ‘executed’ or ‘captured’ as positive words representing a successful candidate

Algorithms

Systemic bias

● Utilising protected attributes including gender, age, race, religion as proxies

Statistical and computational bias

● Simplifying ML models through considering fewer parameters in training datasets ● Selecting the most accurate ML model

● Length of time in prior employment, productivity and number of lost hours ● AI-based hiring systems that claim to glean information about candidates from audio and video ● Focusing too much on the model accuracy results in choosing more male profiles compared to women

Human bias

● Including and or excluding data points in developing ML models

● Amazon hiring algorithms favoured male candidates based on certain key words in their profiles

leads to developing a more transparent AI that human users can understand and better trust (Hutchinson and Mitchell, 2019). AI developers should consider more than just the performance of algorithms based on a given set of metrics. Equally important is the decision-making context (i.e. HRM) in which the AI outcomes are being used (Delecraz et al., 2022). Moreover, the HR function should be careful and aware of the factors being considered by the algorithms and be involved in modifying the inputs used when necessary. For example, to enhance the performance of hiring algorithms, HR professionals should constantly assess whether the factors are job-related (Friedman and McCarthy, 2020). Thus, the designers of machine learning models should be able to adapt to

Mitigating bias in AI-powered HRM  45

Table 4.2

Mitigation techniques

Roles

Mitigation techniques

HR

● Advising the AI development team to have diverse engineers ● Constantly assessing those factors considered in developing algorithms to be HRM related

AI developers

● Developing standards about training datasets and features used in AI models ● Do not only focus on the performance of algorithms based on a given set of metrics and consider the decision-making context (i.e. HRM)

feedback from those who make employment decisions, and HR practitioners should regularly assess the outcomes of the algorithms (Friedman and McCarthy, 2020). The more algorithms are used, the more the need to ensure that any biases will be identified and mitigated. A recent study by Soleimani (2022) explored potential biases in recruitment and selection and investigated how biases were being mitigated in the development of AI recruitment systems (AIRS). Some of the mitigation techniques mentioned above, such as the collaboration of AI developers and HR managers in each phase of developing AIRS, having a diverse AI development team, and managing ML features, are evidenced by this study (Soleimani, 2022). It is clear that both parties have a role to play (Table 4.2). So far, we have explored potential biases in developing AI-HRM systems and some roles and techniques that can assist in mitigation. However, practitioners may encounter challenges when applying these techniques.

CHALLENGES OF APPLYING BIAS MITIGATION TECHNIQUES A fundamental concern remains the lack of diversity in technology (tech) companies, which are largely staffed by male, affluent and white professionals (Smith and Rustagi, 2020). In addition, algorithm development teams comprise technical-focused people who have a very specific, mathematically aligned approach to analysing problems. There are pipeline problems in diversifying engineering teams but it is not impossible to add people with social science and multidisciplinary knowledge to raise issues and help mitigate bias from a nontechnical perspective (Hao, 2019). A second important challenge from the HRM side is that the areas to which AI are applied can be complex to address. According to Paleyes et al. (2022), constantly assessing and monitoring AI models is an open problem, as understating what the key metrics are (i.e. HRM-related factors) is complicated. Selecting the most appropriate feature in developing AI models is time-consuming and requires good domain knowledge (Najafabadi et al., 2015). This can become more challenging, as the size of datasets increases with more prospective features (L’Heureux et al., 2017).

46  Research handbook on human resource management and disruptive technologies

Feature selection is the process to select the most relevant attributes or features in the dataset, and codifying all relevant data can be complicated. For example, it might be thought useful to develop algorithms at work that use comprehensive data about workers’ behaviour, location, performance, emotional states and social relationships as inputs. However, many employee attributes are not codifiable and proxy data can be risky, as noted above, hence the algorithms fail to capture the full lived experiences of workers and the workplace (Meijerink and Bondarouk, 2021). There are also ethical issues around the gathering and deployment of worker data to which employees might not consent or resist (Arrowsmith, 2023). Furthermore, there may be disagreement among HRM (and developer) field experts on what it means to be fair and unbiased, so there is no clear consensus on ‘legitimate’ and ‘irrelevant’ features (Lee, 2019).

CONCLUSION AND RECOMMENDATIONS FOR ACTION While AI can be an enabler of effective HRM, HR and AI developers must maintain a deep understanding of potential biases and errors that may lead to suboptimal and even harmful decisions. This chapter supports awareness of potential biases in developing AI and informs both researchers and practitioners about biased algorithmic decision-making in the HRM context. It also advances HR understanding of how biases occur in developing AI and highlights the respective accountabilities of AI developers and HR professionals in mitigating these. It is recommended that HR advise AI developers to have diversity in their team and to involve nontechnical experts from the HRM domain. AI developers also need to inform HR about the training datasets and features being used in AI models and to constantly evaluate their context and relevance as well as accuracy and performance. This necessitates communication and teamwork skills as well as technical competencies. For their part, professionals within the HR team need to develop the capacity to collaborate with AI developers to design appropriate tools for real-world application.

REFERENCES Albert, E.T. (2019). AI in talent acquisition: a review of AI-applications used in recruitment and selection. Strategic HR Review, 18(5), 215–21. https://​doi​.org/​10​.1108/​shr​-04​-2019​ -0024. Angwin, J., J. Larson, S. Mattu, and L. Kirchner (2016). Machine Bias. ProPublica, 254–64. https://​doi​.org/​10​.1201/​9781003278290​-37. Arrowsmith, J. (2023). Ethics and HRM. In J. Crawshaw, P. Budhwar and A. Davis (Eds.), Human Resource Management: Strategic and International Perspectives. 4th edition. SAGE. Bender, E.M., T. Gebru, A. McMillan-Major and S. Shmitchell (2021). On the dangers of stochastic parrots: Can language models be too big? In FAccT 2021 – Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency (Vol. 1, Issue 1). Association for Computing Machinery. https://​doi​.org/​10​.1145/​3442188​.3445922.

Mitigating bias in AI-powered HRM  47

Berk, R.A. (1983). An introduction to sample selection bias in sociological data. American Sociological Review, 48(3), 386–98. Booth, B.M., L. Hickman, S.K. Subburaj, L. Tay, S.E. Woo and S.K. D’Mello (2021). Bias and fairness in multimodal machine learning: a case study of automated video interviews. In ICMI 2021 – Proceedings of the 2021 International Conference on Multimodal Interaction, 268–77. https://​doi​.org/​10​.1145/​3462244​.3479897. Charlwood, A., and N. Guenole (2022). Can HR adapt to the paradoxes of artificial intelligence? Human Resource Management Journal, December, 1–14. https://​doi​.org/​10​.1111/​ 1748–8583​.12433. Cheng, M., and R. Hackett., (2021). A critical review of algorithms in HRM: definition, theory, and practice. Human Resource Management Review, 31(1), 100698. https://​doi​.org/​ 10​.1016/​j​.hrmr​.2019​.100698. Cowgill, B. (2018). Bias and productivity in humans and algorithms: theory and evidence from résumé screening. Columbia Business School, 29, 1–58. http://​conference​.iza​.org/​ conference​_files/​MacroEcon​_2017/​cowgill​_b8981​.pdf. Cowgill, B., F. Dell’Acqua, S. Deng, D. Hsu, N. Verma and A. Chaintreau (2020). Biased programmers? Or biased data? A field experiment in operationalizing AI ethics. SSRN Electronic Journal, 679–81. https://​doi​.org/​10​.2139/​ssrn​.3615404. Crawford, K. (2013). The hidden biases in big data. HBR Blog Network, 1–4. http://​blogs​.hbr​ .org/​cs/​2013/​04/​the​_hidden​_biases​_in​_big​_data​.html. Datta, A., M. Fredrikson, G. Ko, P. Mardziel and S. Sen (2017). Proxy discrimination in data-driven systems: theory and experiments with machine learnt programs. ArXiv. Delecraz, S., L. Eltarr, M. Becuwe, H. Bouxin, N. Boutin and O. Oullier (2022). Responsible artificial intelligence in human resources technology: an innovative inclusive and fair by design matching algorithm for job recruitment purposes. Journal of Responsible Technology, 11(November 2021), 100041. https://​doi​.org/​10​.1016/​j​.jrt​.2022​.100041. Deshpande, A., N. Picken, L. Kunertova, A. De Silva, G. Lanfredi and J. Hofman (2021). Improving working conditions using artificial intelligence. In Policy Department for Economic, Scientific and Quality of Life Policies, European Parliament (Issue June). http://​ www​.europarl​.europa​.eu/​supporting​-analyses. Eurofound. (2022). Ethics in the Digital Workplace. Publications Office of the European Union. European Parliament (2020). The Ethics of Artificial Intelligence Issues and Initiatives: Study Panel for the Future of Science and Technology. https://​www​.europarl​.europa​.eu/​stoa/​en/​ document/​EPRS​_STU(2020)634452. EU-OSHA.(2022). Artificial Intelligence for Worker Management: An Overview. Publications Office of the European Union. https://​doi​.org/​10​.2802/​76354. Feathers, T. (2021). Major universities are using race as a ‘high impact predictor’ of student success. The Markup, 268–73. https://​doi​.org/​10​.1201/​9781003278290–39. Forde, J.Z., A.F. Cooper, K. Kwegyir-Aggrey, C. De Sa and M. Littman (2021). Model selection’s disparate impact in real-world deep learning applications. ArXiv, 1–8. http://​arxiv​ .org/​abs/​2104​.00606. Friedman, G. and T. McCarthy (2020). Employment law red flags in the use of artificial intelligence in hiring. American Bar Association, September. https://​www​.americanbar​.org/​ groups/​business​_law/​publications/​blt/​2020/​10/​ai​-in​-hiring/​. Fuller, J.B., M. Raman, E. Sage-Gavin and K. Hines (2021). Hidden workers: untapped talent. https://​www​.hbs​.edu/​ris/​Publication Files/hiddenworkers09032021_Fuller_white_pap er_33a2047f-41dd-47b1–9a8d-bd08cf3bfa94.pdf. Hao, K. (2019). This is how AI bias really happens – and why it’s so hard to fix. MIT Technology Review. https://​www​.technologyreview​.com/​s/​612876/​ this-is-how-ai-bia s-really-happensand-why-its-so-hard-to-fix/?utm_source=newsletters&utm_medium=email&utm_campaign=the_algorithm.unpaid. engagement.

48  Research handbook on human resource management and disruptive technologies

Harlan, E., and O. Schnuck (2021). Objective or biased: on the questionable use of Artificial Intelligence for job applications. https://​interaktiv​.br​.de/​ki​-bewerbung/​en/​?utm​_keyword​=​ referral​_input. Houser, K.A. (2019). Can AI solve the diversity problem in the tech industry? Mitigating noise and bias in employment decision-making. Stanford Technology Law Review, 22(2), 291–353. Hüllmann, J.A., S. Krebber and P. Troglauer (2021). The IT artifact in people analytics: reviewing tools to understand a nascent field. International Conference on Wirtschaftsinformatik, 238–254. http://​www​.springer​.com/​series/​11237. Hutchinson, B., and M. Mitchell (2019). 50 Years of Test (Un)fairness. Fairness, Accountability, and Transparency, 49–58. https://​doi​.org/​10​.1145/​3287560​.3287600. IBM. (2018). IBM Watson Talent Frameworks. https://​www​.ibm​.com/​talent​-management/​hr​ -solutions/​ibm​- watson-talent-frameworks. Jo, E.S. and T. Gebru, (2020). Lessons from archives: Strategies for collecting sociocultural data in machine learning. In FAT* 2020 – Proceedings of the 2020 Conference on Fairness, Accountability, and Transparency, 306–16. https://​doi​.org/​10​.1145/​3351095​.3372829 Jordan, M., and T. Mitchell (2015). Machine learning: trends, perspectives, and prospects. Science, 349(6245). Kamiran, F., and T. Calders (2012). Data preprocessing techniques for classification without discrimination. Knowledge and Information Systems, 33(1). https://​doi​.org/​10​ .1007/​s10115–011–0463–8. Kellogg, K.C., M.A. Valentine and A. Christin (2020). Algorithms at work: the new contested terrain of control. Academy of Management Annals, 14(1), 366–410. https://​doi​.org/​10​ .5465/​annals​.2018​.0174. Kirimi, J., and C. Moturi (2016). Application of data mining classification in employee performance prediction. International Journal of Computer Applications, 146(7), 28–35. https://​ doi​.org/​10​.5120/​ijca2016910883. Kleinberg, J., J. Ludwig, S. Mullainathan and A. Rambachan (2018). Advances in big data research in economics: algorithmic fairness. AEA Papers and Proceedings, 108, 22–27. https://​doi​.org/​10​.1257/​pandp​.20181018. Lebovitz, S., N. Levina and H. Lifshitz-Assaf (2021). Is AI ground truth really true? The dangers of training and evaluating AI tools based on experts’ know-what. MIS Quarterly: Management Information Systems, 45(3), 1501–25. https://​doi​.org/​10​.25300/​MISQ/​2021/​ 16564. Lee, M. (2019). Challenges of fairness in AI decisions. Deloitte. https://​ukfinan​cialservic​ esinsights​.deloitte​.com/​post/​102ftmq/​challenges​-of​-fairness​-in​-ai​-decisions. L’Heureux, A., K. Grolinger, H.F. Elyamany and M.A.M. Capretz (2017). Machine Learning with Big Data: Challenges and Approaches. IEEE Access, 5, 7776–97. https://​doi​.org/​10​ .1109/​ACCESS​.2017​.2696365. Lones, M.A. (2021). How to avoid machine learning pitfalls: a guide for academic researchers. http://​arxiv​.org/​abs/​2108​.02497. Lukashin, A., M. Popov, D. Timofeev and I. Mikhalev (2021). Employee performance analytics approach based on anomaly detection in user activity. In Proceedings of International Scientific Conference on Telecommunications, Computing and Control, 321–31. http://​ www​.springer​.com/​series/​8767. Malik, M. (2020). A hierarchy of limitations in machine learning. ArXiv, 1–68. http://​arxiv​ .org/​abs/​2002​.05193. Manyika, J., J. Silberg and B. Presten (2019). What do we do about the biases in AI? Harvard Business Review, 2–6. Margetis, G., S. Ntoa, M. Antona and C. Stephanidis (2021). Human-centered design of artificial intelligence. In G. Salvendy and W. Karwowski (Eds.), Handbook of Human Factors

Mitigating bias in AI-powered HRM  49

and Ergonomics (Fifth, Vol. 19, Issue 3, p. 248). John Wiley & Sons. https://​doi​.org/​10​ .1016/​0003​-6870(88)90184​-6. Meijerink, J., T. and Bondarouk (2021). The duality of algorithmic management: toward a research agenda on HRM algorithms, autonomy and value creation. Human Resource Management Review, 33(1), 100876. https://​doi​.org/​10​.1016/​j​.hrmr​.2021​.100876. Moore, P. (2020). Data Subjects, Digital Surveillance, AI and the Future of Work: Study (Issue December). EPRS | European Parliamentary Research Service. Najafabadi, M.M., F. Villanustre, T.M. Khoshgoftaar, N. Seliya, R. Wald and E. Muharemagic (2015). Deep learning applications and challenges in big data analytics. Journal of Big Data, 2(1), 1–21. https://​doi​.org/​10​.1186/​s40537​-014​-0007​-7. Ntoutsi, E., P. Fafalios, U. Gadiraju, V. Iosifidis, W. Nejdl, M.E. Vidal, S. Ruggieri, F. Turini, S. Papadopoulos, E. Krasanakis, I. Kompatsiaris, K. Kinder-Kurlanda, C. Wagner, F. Karimi, M. Fernandez, H. Alani, B. Berendt, T. Kruegel, C. Heinze, … S. Staab (2020). Bias in data-driven artificial intelligence systems – an introductory survey. Wiley Interdisciplinary Reviews: Data Mining and Knowledge Discovery, 10(3), 1–14. https://​doi​ .org/​10​.1002/​widm​.1356. Paleyes, A., R.-G. Urma and N.D. Lawrence (2022). Challenges in Deploying Machine Learning: a Survey of Case Studies. ACM Computing Surveys. https://​doi​.org/​10​.1145/​ 3533378. Pedreschi, D., S. Ruggieri and F. Turini (2008). Discrimination-aware data mining. In Proceedings of the 14th Acm Sigkdd International Conference on Knowledge Discovery and Data Mining, 560–8. Schwartz, R., A. Vassilev, K. Greene, L. Perine, A. Burt and P. Hall (2022). Towards a Standard for Identifying and Managing Bias in Artificial Intelligence. National Institute of Standards and Technology, 1270. Slota, S.C., K.R. Fleischmann, S. Greenberg, N. Verma, B. Cummings, L. Li and C. Shenefiel (2021). Many hands make many fingers to point: challenges in creating accountable AI. AI and Society, 0123456789. https://​doi​.org/​10​.1007/​s00146–021–01302–0. Smith, G., and I. Rustagi (2020). Mitigating Bias in Artificial Intelligence: An Equity Fluent Leadership Playbook. University of California (Berkeley), Center for Equity, Gender & Leadership Soleimani, M. (2022). Developing Unbiased Artificial Intelligence in Recruitment Framework and Selection : Processual. Massey University. Strohmeier, S. (2022). HR machine learning – an introduction. In S. Strohmeier (Ed.), Handbook of Research on Artificial Intelligence in Human Resource Management (Issue Ml, pp. 25–45). Strohmeier, S. and F. Piazza (2015). Artificial intelligence techniques in human resource management – a conceptual exploration. In C. Kahraman and S.C. Onar (Eds.), Intelligent Techniques in Engineering Management (pp. 149–72). Springer. Sun, W., O. Nasraoui and P. Shafto (2020). Evolution and impact of bias in human and machine learning algorithm interaction. PLoS ONE, 15(8). https://​doi​.org/​10​.1371/​journal​ .pone​.0235502. Tambe, P., P. Cappelli and V. Yakubovich (2019). Artificial intelligence in human resources management: challenges and a path forward. California Management Review, 000812561986791. https://​doi​.org/​10​.1177/​0008125619867910. Tchakounté, F. and F. Hayata (2017). Supervised learning-based detection of malware on android. In M.-H. Au and K.-K.C. Raymond (Eds.), Mobile Security and Privacy (pp. 101–54). Syngress. Terrell, J., A. Kofink, J. Middleton, C. Rainear, E. Murphy-Hill, C. Parnin and J. Stallings (2017). Gender differences and bias in open source: pull request acceptance of women versus men. PeerJ Computer Science, 2017(5). https://​doi​.org/​10​.7717/​peerj​-cs​.111.

50  Research handbook on human resource management and disruptive technologies

World Economic Forum (2021). Human-Centred Artificial Intelligence for Human Resources: A Toolkit for Human Resources Professionals (Issue December). Yapo, A., and J. Weiss (2018). Ethical implications of bias in machine learning. In Proceedings of the Annual Hawaii International Conference on System Sciences, 2018-January, 5365–72. https://​doi​.org/​10​.24251/​hicss​.2018​.668.

5. Navigating through ethical dilemmas, human rights and digital governance Jesús Salgado-Criado and Celia Fernández-Aller

INTRODUCTION Technological solutions require ethical reflection on which is the best option among those available. However, we are so overwhelmed with competing crises and divergent arguments from different interest groups in favor of different solutions that it is often difficult for us to think about what is best among the available options. Carl Mitcham (2012)

Human resources management (HRM) is increasingly using artificial intelligence (AI) in different areas of activity: recruitment, selection, onboarding new employees, training, advancement, retention and employee benefits. Some applications can be the prediction of turnover with artificial neural networks, candidate search with knowledge-based search engines, staff rostering with genetic algorithms, HR sentiment analysis with text mining, résumé data acquisition with information extraction and employee self-service with interactive voice response (Buzko et al., 2016). These may bring many advantages but also may create ethical dilemmas and impacts on human rights, which can result in lower performance and counterproductive work behaviors. The whole reflection is based in a conceptual framework: the human rights approach. As has been stated before (Salgado-Criado, Fernández Aller, 2021), the human rights approach provides a globally accepted structure of principles from which to derive specific regulation. The new European regulation uses a risk-based framework and AI applications are categorized by risk: unacceptable risk, high risk, limited risk and minimal risk. Many applications fall into the last category, and therefore, out of the scope of the proposed legal framework. This approach focuses almost entirely on developers of AI, with almost nothing concerning end-users.1 It is particularly significant that the proposal for a regulation laying down harmonized rules on artificial intelligence (European Commission, 2021) in Europe states that AI systems at work are considered ‘high risk’ because of their possible appreciable effects on the future employability and livelihood of individuals when they are used for: a) the management of persons in an employment relationship, also with respect to self-employed workers, particularly for the selection of persons; b) decision-making regarding promotion or termination of employment or for the distribution of tasks; c) monitoring or evaluation of persons subject to an employment relationship. 51

52  Research handbook on human resource management and disruptive technologies

A rights-based approach would impose the obligation of considering human rights principles as guidance: nondiscrimination, universality of rights, participation of all actors, sustainability, transparency and accountability; the orientation is completely different. How could we operationalize these principles into our organizations? As regards participation, a machine learning project is more likely to be legally compliant and perceived as ethically appropriate if employees are included as stakeholders in the process; results of using AI in human resources must be nondiscriminatory; employers have to offer employees enough information about the rationales for using ML in terms of how it will help them and the organization; furthermore, they have to explain the process that was followed to ensure legality and protect employees’ privacy and other safeguards to protect the security of their data and opportunities for appeals (Hamilton and Davison, 2022). This approach is also based in ethics. There are many ethical codes and abundant literature concerning the ethical dilemmas that arise from AI (Fjeld et al., 2020). Ethical considerations are important, but are not enforceable, and do not enjoy universal consensus, nor universal principles. Human rights are enforceable (through national and international regulation), have a wide consensus and are universally applicable. Additionally, some organizations are using codes of ethics as ‘bluewashing’ (Floridi, 2019).

ALGORITHMS AND HUMAN RESOURCES: IMPACTS ON HUMAN RIGHTS Increasingly, public- and private-sector employers are turning to AI and other disruptive technologies to help with the hiring process, for at least two reasons. The first is capacity: the number of applicants per position has multiplied in the last several years, while staffing levels at human resources (HR) departments remain flat. The second is fairness: there is a growing awareness that hiring processes are rife with implicit bias and discrimination and that hiring decisions often boil down to ‘is this person like me?’ Many organizations believe that AI may offer at least a partial solution to this challenge (Raso et al., 2018). In several studies, researchers have shown that we are less likely to believe in algorithms than in human judgment, especially if we have seen them err (Dietvorst et al., 2015). This is called algorithm aversion. The authors performed an experiment: when the participants were allowed to modify the algorithms to a slight degree, they were more likely to be satisfied with the predictions of the algorithm and, accordingly, more likely to use the algorithm in the future. But the interesting part is the degree to which those modifications could occur. Even if those modifications were insignificant, participants were more likely to believe in the modified algorithms than in the original ones. There are many examples of public employment services using algorithmic profiling models to predict a jobseeker’s probability of finding work, trying to reduce costs and improve efficiency; for example, the Austrian ArbeitsMarktService

Navigating through ethical dilemmas, human rights and digital governance  53

(AMS), a public employment service (PES). AMS automates the profiling of jobseekers to make its counseling process more efficient and to improve the effectiveness of active labor market programs. Based on statistics from previous years, the system calculates the future chances of jobseekers on the labor market using the computer-generated reintegration chance indicator. The algorithmic system looks for correlations between successful employment and jobseeker characteristics, including age, ethnicity, gender, education, care obligations and health impairments as well as past employment, contacts with the AMS and the state of the labor market in the jobseeker’s place of residence. Different levels of assistance and resources for further education become available to the diverse categories of jobseekers (Allhutter et al., 2020). This algorithm has been criticized due to the perceived discriminatory elements within the algorithm, with specific regard to women over 50. The responsibility of organizations to respect human rights applies not just to the services they provide and the products they sell, but also to their internal operations. Flawed hiring processes may have significant implications for the right to freedom from discrimination, the right to equal pay for equal work and the rights to freedom of expression and association. Governments have recognized the need for mechanisms to provide remedy for individuals subjected to discriminatory hiring practices and have created institutions such as the US Equal Employment Opportunity Commission (EEOC) and the Canadian Human Rights Commission. As AI-based hiring systems become commonplace, it will be important to evaluate whether these existing mechanisms are up to the task of ensuring that these new technologies are free from bias. Impacts on Privacy Especially if they share their data with employers, AI-based hiring systems that canvass a wide variety of data sources can impact the privacy of jobseekers and of employees too. HRM information systems use lots of personal data. The information in those systems requires protection, but many companies acting as controllers – that is, the actor who decides the use of personal data – are not aware of how the information is collected, whether or not it is shared with other entities and what it is used for. They simply use the information without taking into account legal requirements. If the companies are European, the GDPR (General Data Protection Regulation) is applicable. The EU General Data Protection Regulation (GDPR) defines ‘processing’ as ‘any operation or set of operations which is performed upon personal data or sets of personal data, whether or not by automated means, such as collection, recording, organization, structuring, storage, adaptation or alteration, retrieval, consultation, use, disclosure by transmission, dissemination or otherwise making available, alignment or combination (…) restriction, erasure or destruction’. Article 8 of the Charter of Fundamental Rights of the European Union (2000/C 364/01) explains the concept of ‘Protection of personal data’ as follows: ‘(1)

54  Research handbook on human resource management and disruptive technologies

Everyone has the right to the protection of personal data concerning him or her. (2) Such data must be processed fairly for specified purposes and on the basis of the consent of the person concerned or some other legitimate basis laid down by law. Everyone has the right of access to data that has been collected concerning him or her, and the right to have it rectified. (3) Compliance with these rules shall be subject to control by an independent authority’ (European Commission, 2012). The right to the protection of personal data is understood as a guarantee of other fundamental rights, a more specific right within the more general right of personal privacy (Troncoso Reigada, 2010). Informed consent is an important element of data protection for information and communication technology (ICT) systems, as the consent of a data subject (an individual) is often necessary for a third party to legitimately process his/her personal data. To provide informed consent regarding the use of personal data, the citizen must have a clear understanding of how his/her personal information will be used by the system. Another interesting point in the GDPR law includes a requirement to provide the right to an explanation. The data subject (employee) has a legally protected right to ask for an explanation of an algorithm’s decisions that impacts their lives. When HRM functions are subcontracted, the companies receiving employee’s personal data are ‘processors’. They act on behalf of the controllers. Controllers can share data with them without the consent of data subjects (employees). Technologies such as people analytics allow us to anticipate organizational risks: not only by measuring absenteeism, for example, but also by classifying people as potentially conflictive or dangerous to the company’s reputation based on their temperament, which some data scientists claim can be measured by mouse movements (Kieslich et al., 2020). But, how reliable is it to translate mouse movements into a psychological profile? How should we measure a person’s values beyond their tendencies? Younger people are aware of the consequences of their lack of privacy: isn’t it a violation of personal freedom to live in fear of publishing online something that could be detrimental to one’s future career?2 Impact on the Right to Equality There are three main challenges for decision-makers applying the predictions produced by machine learning. The first concerns fairness and legal issues, the second relates to a lack of explainability of the algorithm, and the third to the question of how employees will react to algorithmic decisions. As citizens become aware of the right to equality, fairness turns into the main challenge now for machine learning applications. The human resources context raises numerous issues where fairness matters. One of the most obvious of these is the recognition that any algorithm is likely to be backward looking (Tambe et al., 2019). The presence of past discrimination in the data used to train a hiring algorithm, for example, is likely to lead to a model that may disproportionately select on white

Navigating through ethical dilemmas, human rights and digital governance  55

males if, in the past, white males accounted for most of those rated as high performers, and hence perpetuate existing discrimination. AI tools need to be programmed with the right parameters to put a stop to unconscious bias. In an interconnected and complex world, soft skills such as the ability to work in a team or empathy are increasingly important: how does a preselection algorithm include such skills? Perhaps it cannot be an automatic task and will always require a human being in the loop. Selecting people is basically discriminating in the most neutral sense of the term, because it is the action of separating people into two groups: suitable and unsuitable. The machine itself is not biased, but it uses the labels introduced by the data scientist. So what is happening is that we are transferring to the machine the human biases of the person training the algorithm. Does it make sense to correct the bias artificially so that patterns are not repeated? Impact on the Right to Education Concerning the use of AI for training purposes, we have to ask about what interventions make sense for which individuals and if they improve performance. Machine learning can help identify which kind of employee training will be most effective in maintaining product quality. In some cases, eligibility for a particular training program is decided by an algorithm, which may not choose ‘old employees’ if the algorithm includes a prediction on how long they will stay with the company. The selection made by AI in these cases may affect the right to education, which must be available without discrimination. Impact on the Right to Desirable Work AI-based hiring systems could help improve pay equity if designed and implemented correctly, otherwise, they might exacerbate current pay inequalities. Due to the surge of ‘digital labor platforms’, a new kind of workforce has emerged, called ‘digital platform workers’, which has spread worldwide over the last 10 years. This is a category of workers whose activities are determined and controlled by means of continuous and pervasive interaction with sophisticated algorithms. The novelty of the structural relations generated by digital platforms of this kind has brought to the attention of legislators the need for stronger protection of workers, especially due to the vulnerability that afflicts it (Codagnone et al., 2016). Indeed, the working conditions in this field are usually severe, given the many risks undertaken (e.g. road incidents involving delivery riders), the low level of remuneration and the length and distribution of working-shifts. In recent years we have witnessed the intervention of EU legislation on the transparency of working conditions, with the recent Directive (EU) 1152/2019 and recent concerns expressed by the European Parliament (European Commission, 2019).

56  Research handbook on human resource management and disruptive technologies

Current legislation and ordinary judicial remedies are insufficient to solve a complex matter such as the legal issues concerning platform workers, especially those cases impacting workers’ digital rights. The support of international institutions (OECD Framework for the Classification of AI Systems, 2022) and the community of experts is required in order to assess the risks posed by artificial agents. In the EU context, for example, it is worthwhile to mention an attempt to prepare the ground for the future, with the Union’s ‘Artificial Intelligence Law’ (Floridi et al., 2022). For example, indirect discrimination was caused as a result of the use of algorithms in relations with riders, as revealed in the recent judgment of the Tribunale Ordinario di Bologna of 31 December 2020 (Tribunale Ordinario di Bologna, 2020). In that decision, the Italian Court considered that the existence of indirect discrimination against the plaintiff riders had been established, as it had been shown that the platform scored those who were absent voluntarily in the same way as those who were absent due to a legally protected situation, such as union activity. Specifically, its so-called ‘reliability rate’ was affected, penalizing riders who carried out trade union activities. Besides ordinary legislation and judicial proceedings, a few alternative approaches can be envisioned, suitable to offer solutions that can bring a lasting benefit for the workforce but also for platform providers and for society in general. Technological design The incorporation of ethical values in technological devices has been officially recognized by the EU legislator with article 25 of GDPR, regulating the privacy ‘by design’ and ‘by default’ approaches. In this sense, it could be possible to incorporate the protection of digital rights directly into the algorithm governing the platforms, in order to provide built-in operating mechanisms of trade union negotiation and assistance. Collective bargaining agreements It could be useful to include a binding and specific regulation using collective agreements between union workers and employers’ associations. Of course, in this sense, public institutions play a fundamental role of intermediation in order to avoid abuses. This would be fostered for stakeholders by using an ethics by design approach. Local arrangement and code of conduct Since the services provided by such kinds of platforms are strongly territorial (e.g. delivery), it could be an opportunity for municipalities to step up and regulate some specific aspects that could improve significantly the quality of jobs of platform workers (e.g. creating stations or offering shed zones for riders waiting for a call) or assisting workers and employers in adopting voluntary codes of conduct that could not only improve the quality of jobs, or raise the productivity of platforms, but also provide benefit for the whole community (Carta di Bologna) (Municipality of Bologna, 2018).

Navigating through ethical dilemmas, human rights and digital governance  57

Regulatory sandboxes and living labs The concept of a ‘regulatory sandbox’ is included in the EU proposal called ‘Artificial Intelligence Act’ (Artificial Intelligence Act, 2021, articles 53 and 54), which intend to provide regulation concerning the processing of data and the security measures to safeguard the deployment of artificial agents. In this sense, this tool could be used to establish provisional legal frameworks in order to experiment with new forms of regulations and models of interaction suitable to protect the digital rights of platform workers. Impact on the Right of Peaceful Assembly and Association Since all data could be hiring data for AI, people may be terrified of saying certain things or from associating with certain others for fear of the impact on their employability. Involvement in certain organizations, such as ethnic affinity groups or LGBTQ networks, can negatively impact job prospects. AI is actually being used to screen social networks. If people share motivations and sympathy for certain ideas, it may be used against them. This could also impact the right to freedom of expression.

DIGITAL GOVERNANCE The following section will list some digital governance concepts, principles and policies that organizations are designing and adopting in order to manage the ethical risks and dilemmas pointed out previously in this chapter. Motivations for Governance Frameworks in Organizations There might be many reasons why governance as a strategic concept is introduced in a specific activity or domain, but these institutional developments normally respond to two broad objectives: avoiding risks and drawing upon opportunities that affect the organization’s ability to achieve its goals. Therefore, in speaking about new disruptive technologies in the workplace, these are primarily the two questions we should be posing ourselves when talking about governance: ● What risks should we avoid when adopting these technologies and how? ● How can we use these technologies to their full potential for the organization’s goals? We have mentioned above some ethical and legal compliance risks; other risks like reputational, operational or financial risks (e.g. risks affecting organization’s assets, both tangible like profits or intangible like brand or values) will also be operating. In some organizations, principles and values that guide the company’s ethics are also

58  Research handbook on human resource management and disruptive technologies

considered important intangible assets that need protection. All these risks are very context sensitive depending on the organization’s activity. On the positive side, the opportunities also deserve ethical considerations along with other business considerations. Moral considerations will appear when considering the best use of available technology for the professional and personal flourishing of employees that would optimally be aligned with financial and operational goals. Defining Governance As stated in the Introduction to the Handbook on Theories of Governance (Ansell and Torfing, 2022), the term ‘governance’ is one of the most fashionable and frequently used in social sciences. Part of the attraction is a weakening of a centric-hierarchical view of power in political entities and organizations, and the self-perception of multiple actors as resources and instruments for co-produced policy making, instead of reducing them to passive targets and subjects of rules and norms emanating from a center of power, be it a public administration or a CEO. Still, and maybe precisely because of its popularity, the term has a very elusive generic meaning, and that is the reason why it frequently comes with a qualifying prefix. ‘Corporate governance’ normally refers to the institutionalized interaction among the many players – including shareholders, management, boards of directors, employees, customers, financial institutions, regulators and the community at large – involved in the process of directing and controlling private firms (Ansell and Torfing, 2022). Corporate governance regulation, in practice, depends very much on business traditions and regulations. In the USA, the main focus of corporate governance in for-profit organizations is still concentrated in the agent conflicts of owners versus management and controlling owners versus noncontrolling owners (Coates and Srinivasan, 2018), therefore. concentrating on shareholder rights and the fiduciary duties of boards of directors and management. European tradition of corporate governance promotes, especially in Germany and the Nordic countries, the inclusion of employees as components of the board of directors. Globally, the strategic nature of corporate governance very seldom leaves room for operational HRM decisions, with the exception of executive C-level decisions. Nevertheless, board mission statements are increasingly claiming a stakeholder’s perspective, albeit without inviting most stakeholders to the table. Whether this multistakeholder approach is taken seriously in specific corporations or just paying lip service to investors, customers or employees is another matter. The publication of ISO 30408 ‘Human resource management – Guidelines on Human Governance’ has given rise to the concept of ‘HR governance’ as part of corporate governance (International Standards Organization, 2016). However, due to the still vague definition of both concepts, it is still unclear how they interrelate amongst themselves and with HR management (Kaehler and Grundei, 2018). In the digital domain, the ‘governance’ realm has been taken onboard with enthusiasm: ‘IT governance’, ‘data governance’, ‘information governance’, ‘record and information management’, ‘algorithmic governance’ and other similar terms have

Navigating through ethical dilemmas, human rights and digital governance  59

been used with profusion and, many times, interchangeably. We will not try to clarify their claimed differences here, since faithful conceptual differences are often mixed with branding differentiation motivations. For our purposes, digital governance will be defined as a subordinate to corporate governance: we are interested in the role of the digital resources (data and processes) within corporate governance and their effective control mechanisms in the aspects that might impact the workspace and specifically within the human resources functions in the organization. Digital assets are data of many kinds (structured and unstructured, numeric, textual and multimedia), stored digitally, uniquely identifiable, which organizations can use to realize value. Examples of digital assets include documents, audio, video, logos, slide presentations, spreadsheets and websites (Gartner Group, 2022). Floridi (2018) describes three different but overlapping domains of activity when thinking about the digital facets of the company: digital governance, digital ethics and digital regulation (see Figure 5.1).

Source:

Adapted from Floridi (2018).

Figure 5.1

Three digital facets of organization

Digital governance is a managerial practice ‘of establishing and implementing policies, procedures, and standards for the proper development, use and management of the infosphere’ (Floridi, 2018). The ‘infosphere’ is understood as that liminal space we humans currently inhabit in our personal lives and also in our workplace that is not analogue or digital, but both, not offline nor online, but both. Our human

60  Research handbook on human resource management and disruptive technologies

judgment and decision-making (our ‘human algorithms’) coexist and interface in this space with alien digital algorithms. This coexistence is happening effectively in our workplace, and in fact we are currently working along with colleagues and algorithms, or, at least, human interactions are mediated by algorithms. Historically it was people, workers, that got stuff done. Today algorithms and robots are setting the reference with respect to our productivity and often measuring it. Digital ethics informs both digital governance and digital regulation in a normative way of how to transform ethical values and principles into the infosphere. Digital regulation imposes some legal limits to the organization’s usage of the digital. Digital governance exceeds the domains of digital ethics and digital regulation since there are areas of digital decisions that might not be ethically or legally meaningful. For example, ethical or legal considerations may be irrelevant in technical decisions involving the efficiency of a model or an algorithm. Human Resources and Digital Governance Digital resources and processes of any sort (marketing, finance, operations) have an impact on employees. So, HRM should not only monitor those digital applications used for specific HR functions (sourcing, recruiting, onboarding, talent development, etc.) but all digital applications that, directly or indirectly, may have an impact in the workplace. Digital processes are currently influencing and intermediating many operational functions, coordination and communication within the organization and, therefore, influencing productivity and the wellbeing of employees. Aneesh (2009) and Danaher (2016) adopt the term ‘algocracy’ to indicate a governance system where decisions and even definitions of rules are produced or influenced by the results of computer algorithms. It should be noted that algocracy can partially coexist with other bureaucratic governance systems. It might appear somewhat futuristic, but it is already today very frequent in many industries and corporate functions: decisions about prices in the airlines or hospitality industries have long been automated to algorithms; store opening and closure decisions are also taken by algorithms in big retail chains. Even if these decisions are not as controversial as those that involve the classification of persons or the prediction of people’s behavior, their consequences may impact the whole business and, therefore, the lives of employees, providers and other stakeholders. As noted before, algorithms and humans coexist in the organizational system. The operational environment defines the participation of employees and algorithms as components of the overall organizational system. In order to get a sense of what a hybrid human-algorithmic organizational governance may look like, it can be useful to see a specific function from a command and control perspective. Using a causal graph borrowed from systems thinking and automatic regulation, we could basically depict an organization, or a function within the organization, in terms of

Navigating through ethical dilemmas, human rights and digital governance  61

a generic control loop. Conventional control theory would distinguish between the several components of a control or management system (see Figure 5.2): ● The operational process that produces the output that the organization wants to manage. Say for instance a sales process. ● Feedback systems are all data and knowledge gathering systems trying to obtain all relevant variables that can estimate the internal state of the business process. In our example, sales data reported to management functions. More and more, these feedback systems are digitized, and have less human intervention in its operation. This digitization pursues a more reliable, fast and frequent data gathering. This enables the complete system to be able to manage phenomena that have faster dynamics. ● Set points or goals of the organization established for the operations. In our case the organization may set sales objectives. These set points are compared to the output of the feedback systems. These are then interpreted as ‘deviations’ from the goal. ● The controller, in view of these deviations and with knowledge about the process dynamics (in our case knowledge about sales team skills, company’s offering and market demand), sets the command signals to the process (for instance, product discounts or sales team incentives) in order to correct the internal state and output of the organizational process. Of course, this basic command and control loop is normally embedded within a network of informal human interactions, which would need to be taken into account in a specific team and could eventually cause the hybrid processes to drift from the initial dynamic.

Figure 5.2

Management control system

Conventional management uses human judgment in the ‘controller’ part and the process is normally a complex mix of humans and technology embedded in the business models. With respect to the (partial or total) automatization of the different components in the feedback loop and borrowing from military terminology, humans could be ‘in the loop’, ‘on the loop’ or ‘out of the loop’. Human in the loop refers to the situation where a human completely controls the controller output; in the sales

62  Research handbook on human resource management and disruptive technologies

example, the sales manager deciding on, for instance, a price promotion. Human on the loop refers to the situation where the algorithm is by default making automatic decisions and humans have an oversight function, having the capability to start and end the automatic decision process of the algorithm; in our case, this would occur in dynamic pricing applications for ecommerce. Human out of the loop would mean a completely automated environment where algorithms perform the decision making and also the oversight function. The two latter configurations are recommended in very fast dynamics where time is of the essence (like, for instance, a preselection process on a massively responded-to job posting or route assignment in a logistic network). When decisions have a critical impact for the organization and the time frame is not as stringent, then a human in the loop would be recommended. Human in the loop (both in the ‘controller’ and in the ‘feedback’ processes) is especially recommended where the dynamics are not very fast and also the process has an important human component, like in creative development teams. In some cases, the human in the loop format will turn problematic if the algorithmic response is consistently accurate (like in health diagnostics for some work-related diseases): in these cases the human will tend to avoid opposition to the algorithm, since it will trigger escalation. In any case, the automation options and human involvement for each decision appears to be more like a continuum rather than a binary choice (automate/not automate). See, for instance, the 10 steps for decision automation, adapted from Parasuraman et al. (2000), in Table 5.1. Digital Governance Design Principles Before addressing some of the main issues that workplace and HRM digitalization is posing, we would like to point out some generic governance principles that shape the implementation of digital governance within organizations. Systemic impact of digitalization and human adaptation Every technology, even the most simple and autonomous, has an impact on the organization that is transmitted through a network of social and technical relationships. This transmission potentially affects all functions and components of the organization and these components of the organization (people) often adapt dynamically to a new element, like an algorithm, and its behavior and impact. This can be seen in an example: if recruiting systems have a bias towards those CVs having more words associated with achievements like ‘executed’ or ‘extract value’, people will learn to include those in their CVs or Linkedin profiles, making this a game of zero gain, like the organic positioning in the search engine: trying to game the algorithm, rather than providing relevant information. This already happens in job interviews (still not frequently mediated by an algorithm): Every job candidate knows by now how to answer standard interview questions like ‘What is your worst defect?’ with something like ‘I am too perfectionist.’

Navigating through ethical dilemmas, human rights and digital governance  63

Table 5.1

Levels of automation of decision and action selection

Levels of decision and action selection Algorithm

Human

Lower

  1

No assistance

Human evaluates and takes all

 

2

Offers a complete set of alternatives

Takes all decisions and actions

 

3

Narrows the selection to a few

Takes all decisions and actions.

 

4

Suggest one alternative

 

5

Suggests one alternative and executes Approves suggestion

 

6

decisions and actions

Decides also narrowing criteria. Approves suggestion and executes

if approved Suggests one alternative, allowing

Can veto suggestion for restricted time

restricted time for veto before

before execution

execution    

7 8

Automatically takes decisions and

Receives information on every

executes

decision

Automatically takes decisions and

Receives information on demand on

executes

individual or aggregated decisions and can revert to lower levels of automation

 

9

Automatically takes decisions and

Receives alarms on confidence of

executes

decisions. Can revert at any time to other level or switch algorithm

Higher

10

Complete autonomy

At least for a period of time, does not have any power to revert to other level, veto or switch algorithm.

Source:

Adapted from Parasuraman et al. (2000).

Comprehensible policies One of the main barriers to digital governance is that it might have impact in several organic functions in the organization: IT systems, data management, development, human resources management, marketing and business units. All have their ‘piece’ but digital governance must have a unique direction, recognized by all functions. Silos and power bubbles are the most important barriers to defined, mature digital governance. Instrumental and procedural legitimacy The importance of legitimate decision-making in organizations has been stressed by many. Instrumental legitimacy is provided when decisions are more effective in achieving organizational aims and objectives. Procedural legitimacy is concerned with the righteousness of the process that provides, as for example in Habermas (1990), space to all affected parties to participate. At a minimum, procedural legitimacy includes general awareness, opportunities for reasoned rejection and general acceptance.

64  Research handbook on human resource management and disruptive technologies

Principles for Digital Governance in the Workplace The following is a list of issues that digital governance norms and procedures incorporate with respect to the adoption of digital resources and processes in the workplace. Stakeholders’ alignment One of the main goals of corporate governance (and, therefore, of its digital components) is to align the interests of the different stakeholders. Issues of misalignment are often found, for instance, in the use of some types of ‘digital bartering’, for example, where digital services are offered to users in exchange for personal data. Clearly, company management or investors’ interests in this particular case might not be well aligned with customers’ interests. Internally, these interactions are also conceivable to involve employees. A mobile device, a car with GPS navigator or a ‘coaching bot’ offered to managers include digital resources (among others, the personal data they generate) that should be under the control of digital government policies, and alignment between employees’ and company interests should be an important goal identified in digital governance. Autonomy of decisions and accountability When algorithms take the role of decision-making, humans loose some degree of autonomy in the workplace. If operational decisions are taken by algorithms, employees may feel no longer accountable for the consequences, especially if algorithm decisions are difficult to interpret. Employees – e.g. customer care agents – are often required to explain decisions taken by a model. Take the representative of a bank branch that communicates loan decisions to customers. The alienation experienced by the bank agent of having to communicate a decision that is taken by an algorithm3 and not understood by the employee is within HRM scope. This psychological separation of the employee from the decision and the organization is not solved, but exacerbated, if the ‘last decision’ is assigned to the employee (i.e. introducing ‘human in the loop’). This disconnection with the decision and the organization is more painful if there is a close relationship with the customer (which is what the customer care department is supposed to do). Honest human–digital interaction The type and quality of interactions between employees and digital resources of the company have more and more influence on the productivity and the wellbeing of employees. Many AI applications (not only HR-specific ones) have an interface with users (be they employees, candidates or managers) in the form of bots or voice assistants. Most companies recognize that bots or voice assistants should not deceive the user by impersonating a human, thus making the user think that he or she is talking to another human, and normally the company has a standard statement advising the user of this fact. Nevertheless, according to Paradox.ai, a bot that helps candidates during the recruiting process, in their 2021 report (Paradox Inc., 2021), more than 2 million

Navigating through ethical dilemmas, human rights and digital governance  65

candidates ended their conversations by thanking the bot. This is a sign that users either do not read the statement (clearly shown in the graph) or believe that the thank you message will help them in the recruitment process, or forget the digital nature of the bot and may bond somehow during the conversation with the digital assistant. Even if it is difficult to discern between these different motivations, the confusion between the human or digital nature of the bot is to be avoided (at the expense of engagement) through the digital governance policies through some measures, like, for instance, not using a human photo or a human name identifying the bot.4 Principles for Digital Governance in HRM The following is a list of issues that should be addressed in the norms and processes specified in digital governance of HRM operations. Transparency and explainability At least HRM should be involved in the decisions affecting the employees directly. An employee would, in this sense, have the right to receive an explanation about selection, promotion, salary modification, appraisal and decisions that affect him or her. Privacy ‘Digital exhaust’ provides more authentic data to analyse some common concerns of HRM. Emails or internal messaging that employees use while working are especially handy for HR analysis. It could be arguably said that ‘issue detection’ in HR is the best candidate for this analysis. Examples of detection of such issues would be models to detect potential internal fraud, harassment, resistance to change or ‘flight risk’. Because of the nature of these findings (critical consequences for the employee, but not pressured by time requirements), they suggest a human in the loop in the evaluation and decision making. But here another feedback loop will operate that may affect the organization: when employees become more aware of the use that HRM is making of emails and messaging platforms, the employees will be less forthcoming and more wary while using these tools, and maybe other organizational benefits of these tools will deteriorate. Fair attribution of performance to individuals Most jobs today are a complex network of interrelations and it is difficult to separate individual merits from group merits (Tambe et al., 2019). Therefore, the assessment of individual performance is an area where managers’ judgment is deemed to have prevalence. The mere existence of a data-based insight coupled with managers’ time pressure might have a negative outcome. Contemplation of technical constraints There have been many reported technical constraints to be taken into account when using AI for HRM, like constraints of small data sets (Tambe et al., 2019), that affect

66  Research handbook on human resource management and disruptive technologies

accuracy and false positive and negative errors, the existence of open loops without model monitoring (O’Neil, 2017) (no organization recruits low-ranking candidates to ascertain the performance of the model in type 2 errors – false negatives) and model drift (model performance degrading with time due to group or individual changes overtime). Accountability and the ‘many hands’ problem It is normally considered that whoever makes the decision is responsible for its consequences. But decisions are made based on insights and insights based on technology and data. Decision-makers can disengage from insights and data if they are not responsible for those. In an example, a sales manager makes a hiring decision based on a machine learning model developed by the data science department based on a vendor ML library for HRM, which takes into account some historical financial performance data. As can be seen, several departments are accountable (decision, insights, technology and data). The problem of accountability occurs in the form of ‘many hands’ (Nissenbaum, 1996). A tension between efficiency and accountability will involve organizational decisions: a more efficient centralized data science department that serves all departments or a dedicated data science team in each department. Vigilance of effects of algorithms on people There are several effects that could outcrop from the utilization of data analytics or classifications that need to be taken into account. For instance the ‘Pygmalion effect’: people react to what they understand is the perception of others. If an algorithm tells the manager that a specific employee is, for instance, a leader with peers, then the manager will treat this employee differently from the moment she receives the information. Furthermore, the employee, seeing the boss reacting positively, will feel inclined to ‘deserve’ the opinion. All of which reinforces the behavior of both employee and manager. The same can be said in negative terms: if the algorithm classifies an employee with lack of collaboration, the chain reactions of the manager and the employee will probably end by making correct a self-fulfilling prophecy. Likewise, in terms of self-esteem: if results are shown to all organizations, a weak position in the organization could affect self-esteem of workers if that position is exposed to all. A certain degree of transparency (for instance, the way in which employee data is used and who in the organization has access to the data) but any kind of scoring, can be very harmful for the employee relations within the organization, and hinder productivity. Fairness discussions: human bias vs. algorithmic bias There is a very extensive literature that shows that recruitment via ML can be unfair. In most accounts, the consideration of fair metrics is regarded as non-Boolean but as a continuous variable. In this context, successive optimization of fairness metrics is possible; and the goal of increasing fairness metrics is worthy. Now the reference for ‘good’ versus ‘excellent’ outcomes should come from a nonalgorithmic (i.e. conven-

Navigating through ethical dilemmas, human rights and digital governance  67

tional means of recruiting) source. If the results of the application of ML techniques are proved fairer than the conventional (human not data-based and sometimes based on intuitions) recruitment process, should it not be implemented as soon as possible, and then monitored and optimized? Discussions about which metrics of fairness to choose for each business case are relevant but should not, in most cases, unnecessarily delay implementation. It would just not be fair.5 Systemic and nonsystemic interventions Algorithms may amplify an HR problem that we would like to avoid. For example, consider the discovery of a discriminatory hiring problem in a model that does the preselection of candidates. The inclusion of synthetic data to augment the chances of minority groups as a positive action could be considered as a solution. Now, if the model is discriminatory because it reflects a company culture that is hostile to these minority groups, the modified algorithm will not change the true reason for the discrimination but the symptom. The true problem is the company culture and the modified algorithm would just admit more minority candidates, who will eventually have to leave the company, adding a high-rotation problem to the unsolved discriminatory problem in the organization. The job is to dismantle this culture and a hiring algorithm will not do the job. This is an example of the danger of a nonsystemic intervention when a systemic one is required. The archetype ‘shifting the burden to the intervenor’ used by system thinking is the one operating here, where algorithmic decision-making is taking the place of ‘the intervenor’ (Senge, 2006). Avoidance to coding the status quo Self-fulfilling prophecies is another effect that decisions based on ML models may cause if the complete picture is not regarded in their design. For example, gender is in most companies correlated with voluntary turnover prediction. A model that recommends training based on turnover prediction will select men more frequently, which will probably result in women’s disengagement and turnover. In this case, algorithm synthetic de-biasing techniques may actually be effective in changing a situation that the organization wants to change. Reliability and applicability of external data Sometimes, because the data sets available to the company are too small, the recourse to external data from vendors or consultants may seem a good option. Of course, not all working contexts and company cultures are similar and they do have an impact on the model accuracy.

REGULATORY EFFORTS AND PROFESSIONAL STANDARDS Several regulatory developments are being initiated in the EU, US, China and other parts of the world. Multilateral bodies are also embracing these regulatory efforts.

68  Research handbook on human resource management and disruptive technologies

Although in most cases the regulatory processes are still in their early days, we can already provide some general ideas on their general contents and approaches in HRM and workplace impacts. While some voices are still suspicious of regulating AI, the idea that regulation is required to set a minimum standard for AI responsible design and use is finally reaching consensus in most countries. Regulating AI is expected to benefit companies that use and develop AI, with clear indications on the limits, which will define a level playing field vis-a-vis competition (especially international competition) and will make design and implementation decisions easier since they will rely less on the always somewhat subjective moral considerations on the part of the companies. At the same time society at large will benefit from more trustworthy applications of technology, provided that regulations put the necessary enforcement mechanisms in place and governments provide enough resources to meet enforcement requirements. The EU has a relevant regulatory tradition when it comes to digital technologies. Its General Data Protection Regulation (GDPR) was in fact a major breakthrough in regulation. The European Commission submitted in April 2021 its draft AI Act (European Commission, 2021), taking a position of specific technology regulation. Once it is passed, these new rules could take some time to take effect, as European legal procedures are slow. EU-proposed regulation on AI6 (European Commission, 2021) defines some high risk applications that should go through more stringent certification and auditing processes in Annex 3.7 Specifically, as stated in the introduction of this chapter, the proposed regulation defines two high risk applications8 in the area of employment, workers management and access to self-employment: (a) AI systems intended to be used for recruitment or selection of natural persons, notably for advertising vacancies, screening or filtering applications, evaluating candidates in the course of interviews or tests; (b) AI intended to be used for making decisions on promotion and termination of work-related contractual relationships, for task allocation and for monitoring and evaluating performance and behavior of persons in such relationships. Surprisingly, the AI Act proposal does not introduce any protection or redress mechanisms for employees that may have their rights damaged. It only imposes obligations on organizations developing or using AI. The US approach to AI regulation has been different. Instead of defining a single regulation on this specific technology, US regulators are making specific amendments to sectorial regulations in order to accommodate the specific impact of AI into each policy area. In the HRM area, the main activity that, for the time being, is being regulated is selection and hiring: in October 2021, the US Equal Employment Opportunity Commission (EEOC, 2021) launched an initiative to ensure that AI used in hiring and other employment decisions complies with federal civil rights laws. Other administrations, like New York City and the District of Columbia have proposed similar initiatives regarding AI usage for workers selection. It is very probable that, with this sectorial approach, the US will see AI regulation being passed and enforced before it happens in Europe, since the regulatory process of the AI Act

Navigating through ethical dilemmas, human rights and digital governance  69

within the European Parliament and European Council could cause the act to come into force in 2026.9 China has issued new regulations on algorithmic recommender systems.10 The regulation provides some articles that are applicable in the HRM area, like article 20, that specifies: ‘Where algorithm recommendation service providers provide workers with job scheduling services, they shall protect workers’ legitimate rights and interests such as labor remuneration, rest and vacation, and establish and improve relevant algorithms for platform order distribution, remuneration composition and payment, working hours, rewards and punishments, etc.’ (Provisions on the administration of algorithm recommendations for internet information services, 2022). These are still early days for AI regulation, and it will no doubt be an area of very fast evolution. The EU–US Trade and Technology Council (TTC) is promoting a convergence of both regulations. A more unified international approach to AI governance would be desirable, and could strengthen common oversight, guide research into shared challenges and promote the sharing of best practices, code and data. A global approach to regulation is needed, and a human rights approach would be the way to develop common minimum requirements for AI. This is required since this technology can be accessed by anyone via the cloud.11

PROFESSIONAL AND BUSINESS PRACTICES In parallel to regulatory developments, soft governance activity has proliferated worldwide. Standardization bodies like the International Standards Association (ISO), professional associations like the Institute of Electrical and Electronics Engineers (IEEE) and Technology companies like IBM and Microsoft have proposed similar frameworks to guide and evaluate AI developments in various contexts, including human resources management. These standards, mainly developed as recommended practices and guides, are mainly directed towards the areas of data and algorithmic governance. For instance, IEEE has recently published their ‘Standard for Transparent Employer Data Governance’, which defines specific methodologies to help employers to certify how they approach accessing, collecting, storing, utilizing, sharing, and destroying employee data (IEEE, 2021). The standard provides specific metrics and conformance criteria regarding these types of uses from trusted global partners and how vendors and employers can meet them. Furthermore, some proprietary frameworks, like the IT governance framework Control Objectives for Information Technologies (COBIT) from the Information Systems Audit and Control Association (ISACA), that have a long tradition in information systems auditing, are also updating the framework to include AI technologies and impact (ISACA, 2018).

70  Research handbook on human resource management and disruptive technologies

CONCLUSION AND RECOMMENDATIONS FOR ACTION New digital technologies are continuously disrupting the way we live and work. Our organizations are adopting these technologies mainly because of their promise. The adoption of such technologies needs careful consideration in terms of corporate goals and principles. The digital resources are affecting many activities that before were only performed by employees in different company functions with operational or strategic scope. Machines are mediating, supporting or directly substituting humans in their perception and analysis of the environment, their decisions and actions affecting the organization’s assets and resources. One of the most critical situations we need to analyze is when these resources upon which decisions are taken or recommended are the human resources themselves. Human resources management should closely examine all these and discern positive and negative impacts to employees. Much has yet to be said and researched about how we should guide and design our organizations in this digitalization process. Ethical guides and principles and legal norms are being currently defined so human rights are protected, as a minimum, in this global challenge. Governance policies are also being drawn so we not only responsibly limit the risks for people and assets but gain from these technologies to their full potential for human flourishing.

NOTES 1. In the EU AI act, the ‘user’ is defined as ‘any natural or legal person, public authority, agency or other body using an AI system under its authority, except where the AI system is used in the course of a personal non-professional activity’ so end-users or citizens at large (consumers, employees) are almost not considered in the proposal framework. 2. https://​ w ww​ . itd​ . upm​ . es/​ 2 020/​ 1 1/​ 1 1/​ s emana​ - de ​ - la ​ - ciencia ​ - el ​ - futuro ​ - digital ​ - que​ -queremos​-depende​-de​-lo​-que​-hagamos​-en​-el​-presente/​. 3. Or, for the sake of a general discussion, when decisions are taken from an alien agent (that could be human) but the person in charge of communicating is not provided with enough information about the reasons this decision was taken. In the case of some ML models, the reasons might not even be accessible at this point. 4. The anthropomorphization of an application interface is a significant challenge in Generative AI, where Large Language Models are trained on human-written text that often expresses emotions. If not carefully adapted, these models can inadvertently project these emotions onto themselves, potentially leading to user confusion. 5. In the context of Generative AI, the focus of fairness discussions does not lie primarily on resource disparities but on the potential harms associated with representation, such as the propagation of stereotypes embedded within the original training corpus. 6. The proposal has been subject to trilateral negotiations involving the European Commission (responsible for the first draft), the Parliament and the Council of the European Union. On 9 December 2023, the European Parliament reached a provisional agreement with the Council on the Artificial Intelligence Act. At the time of writing of this chapter, the agreed text is expected to be formally adopted by both Parliament and Council to become EU law during 2024.

Navigating through ethical dilemmas, human rights and digital governance  71

7. Before placing a high-risk AI system on the EU market or otherwise putting it into service, providers must subject it to a conformity assessment. Providers of high-risk AI systems will also have to implement quality and risk management systems to ensure their compliance with the new requirements and minimise risks for users and affected persons, even after a product is placed on the market. In case of a breach, the requirements will allow national authorities to have access to the information needed to investigate whether the use of the AI system complied with the law. 8. The AI Act considers as unacceptable risk a very limited set of particularly harmful uses of AI that contravene EU values because they violate fundamental rights and will therefore be banned. For example, emotion recognition in the workplace and education institutions, unless for medical or safety reasons (i.e. monitoring the tiredness levels of a pilot). 9. In addition to this, an Executive Order of the President has been recently passed: ‘Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence,’ 88 Federal Register 75191, 1 November 2023, at https://​ www​ .federalregister​ .gov/​ documents/​ 2023/​11/​01/​2023​-24283/​safe​-secure​-and​-trustworthy​-development​-and​-use​-of​-artificial​ -intelligence. Sec. 6 establishes measures to support workers in relation to the introduction of AI. Other important references are: White House, Blueprint for an AI Bill of Rights: Making Automated Systems Work for the American People, October 2022, at https://​www​.whitehouse​.gov/​wp​-content/​uploads/​2022/​10/​Blueprint​-for​-an​-AI​-Bill​-of​ -Rights​.pdf; and National Institute of Standards and Technology, Artificial Intelligence Risk Management Framework (NIST AI 100-1), January 2023, at https://​nvlpubs​.nist​ .gov/​nistpubs/​ai/​NIST​.AI​.100​-1​.pdf. 10. On 15 August 2023, a new Chinese law designed to regulate generative AI came into force: Interim Measures for the Management of Generative Artificial Intelligence Services. http://​www​.cac​.gov​.cn/​2023​-07/​13/​c​_1690898327029107​.htm. 11. Although there are now efforts in many countries to separate the internet, like in China and Russia, because of political willingness to exert tight controls over citizens’ access to information and opinions.

REFERENCES Allhutter, D., Cech, F., Fischer, F., Grill, G. and Mager, A. (2020). Algorithmic profiling of job seekers in Austria: How austerity politics are made effective. Frontiers in Big Data, 3, 5. Retrieved from https://​www​.ncbi​.nlm​.nih​.gov/​pubmed/​33693380 Aneesh, A. (2009). Global labor: Algocratic modes of organization. Sociological Theory, 27(4), 347–370. doi:10.1111/j.1467–9558.2009.01352.x Ansell, C. and Torfing J. (2022). Introduction to the handbook on theories of governance. Handbook on theories of governance (pp. 1–16) Edward Elgar Publishing. doi:10.4337/9781800371972.00007 Retrieved from https://​www​.elgaronline​.com/​view/​ edcoll/​9781800371965/​9781800371965​.00007​.xml Buzko I., Dyachenko Y., Petrova M., Nenkov N., Tuleninova D. and Koeva K. (2016). Information and computer technologies artificial intelligence technologies in human resource development. Computer Modelling & New Technologie, (20(2) 26–29) Coates, J. and Srinivasan S. (2018). Finance reading: Corporate governance . Codagnone, C., Biagi, F. and Abadie, F. (2016) The Passions and the Interests: Unpacking the ‘Sharing Economy’. Institute for Prospective Technological Studies, JRC Science for Policy Report, 2016, Available at SSRN: https://​ssrn​.com/​abstract​=2​ 793901 or http://​dx​.doi​ .org/​10​.2139/​ssrn​.2793901

72  Research handbook on human resource management and disruptive technologies

Comune di Bologna (2018). Carta dei diritti fondamentali del lavoro digitale nel contesto urbano. Retrieved from http://​www​.comune​.bologna​.it/​sites/​default/​files/​documenti/​ CartaDiritti3105​_web​.pdf Danaher, J. (2016). The threat of algocracy: Reality, resistance and accommodation. Philosophy & Technology, 29(3), 245–268. doi:10.1007/s13347–015–0211–1 Dietvorst, B. J., Simmons, J. P. and Massey, C. (2015). Algorithm aversion: People erroneously avoid algorithms after seeing them err. Journal of Experimental Psychology. General, 144(1), 114–126. doi:10.1037/xge0000033 EEOC (2021). EEOC launches initiative on artificial intelligence and algorithmic fairness. Retrieved from https://​www​.eeoc​.gov/​newsroom/​eeoc​-launches​-initiative​-artificial​ -intelligence​-and​-algorithmic​-fairness. European Commission (2012). Charter of Fundamental Rights of the European Union. 2012/C 326/02. Official Journal of the European Union. 26.10.2012. Available at: https://​eur​-lex​ .europa​.eu/​eli/​treaty/​char​_2012/​oj. European Commission (2019). Directive (EU) 2019/1152 of the European Parliament and of the Council of 20 June 2019 on transparent and predictable working conditions in the European Union. Available: http://​data​.europa​.eu/​eli/​dir/​2019/​1152/​oj. European Commission (2021). A regulation of the European Parliament and of the council laying down harmonised rules on artificial intelligence (artificial intelligence act) and amending certain union legislative acts COM/2021/206 final Available at: https://​eur​-lex​ .europa​.eu/​legal​-content/​EN/​TXT/​?uri​=​celex​%3A52021PC0206. Fjeld, J., Achten, N., Hilligoss, H., Nagy, A. and Srikumar, M. (2020). Principled artificial intelligence: Mapping consensus in ethical and rights-based approaches to principles for AI. ().Berkman Klein Center for Internet& Society. Retrieved from http://​nrs​.harvard​.edu/​ urn​-3:​HUL​.InstRepos:​42160420 Floridi, L. (2018). Soft ethics and the governance of the digital. Philosophy & Technology, 31(1), 1–8. doi:10.1007/s13347–018–0303–9 Floridi, L. (2019). Translating principles into practices of digital ethics: Five risks of being unethical. Philosophy & Technology, 32(2), 185–193. doi:10.1007/s13347–019–00354-x Floridi, L. and Holweg, M., Taddeo, M., Amaya Silva, J., Mökander, J. and Wen, Y. capAI – A Procedure for Conducting Conformity Assessment of AI Systems in Line with the EU Artificial Intelligence Act (March 23, 2022). Available at SSRN: https://​ssrn​.com/​abstract​=​ 4064091 or http://​dx​.doi​.org/​10​.2139/​ssrn​.4064091 Gartner Group (2022). Gartner glossary. Retrieved from https://​www​.gartner​.com/​en/​finance/​ glossary/​digital​-assets Habermas, J. (1990). Discourse ethics - notes on a program of philosophical justification. The communicative ethics controversy (pp. 60–110) Retrieved from https://​search​.proquest​ .com/​docview/​37174533 Hamilton, R. H. and Davison, H. K. (2022). Legal and ethical challenges for HR in machine learning. Employee Responsibilities and Rights Journal, 34(1), 19–39. doi:10.1007/ s10672–021–09377-z IEEE (2021). IEEE standard for transparent employer data governance IEEE. doi:10.1109/ IEEESTD.2021.9618905 International Standards Organization (2016). Human resource management – guidelines on human governance (First ed.) ISACA (2018). COBIT 2019 framework – governance and management objectives. Schaumburg, IL: ISACA. Kaehler, B. and Grundei, J. (2018). HR governance as a part of the corporate governance concept. HR governance (pp. 27–50). Cham: Springer International Publishing. doi:10.1007/978–3-319–94526–2_3 Retrieved from http://​link​.springer​.com/​10​.1007/​ 978–3​-319–94526–2​_3

Navigating through ethical dilemmas, human rights and digital governance  73

Kieslich, P. J., Schoemann, M., Grage, T., et al. (2020). Design factors in mouse-tracking: What makes a difference? Behav Res 52, 317–41. https://​doi​.org/​10​.3758/​s13428​-019​ -01228​-y Mitcham, Carl (2012). Technology and the burden of responsibility. In: Values and Ethics for the 21st Century, p. 149. Available from: https://​www​.bbvaopenmind​.com/​en/​articles/​ technology​-and​-the​-burden​-of​-responsibility/​ Municipality of Bologna (2018). Charter of Fundamental Rights of Digital Labour in the Urban Context (2018b), Available at: http://​www​.comune​.bologna​.it/​sites/​ default/files/ documenti/CartaDiritti3105_web.pdf Nissenbaum, H. (1996). Accountability in a computerized society. Science and Engineering Ethics, 2(1), 25–42. doi:10.1007/BF02639315 OECD (2022), OECD Framework for the Classification of AI systems, OECD Digital Economy Papers, No. 323, OECD Publishing, Paris. Available: https://​doi​.org/​10​.1787/​ cb6d9eca​-en. O’Neil, C. (2017). Weapons of math destruction. London: Penguin Books. Paradox Inc. (2021). Paradox.ai: A year in review. Retrieved from https://​web​.archive​.org/​ web/​20211229183259/​https://​www​.paradox​.ai/​year​-in​-review. Parasuraman, R., Sheridan, T. B. and Wickens, C. D. (2000). A model for types and levels of human interaction with automation. IEEE Transactions on Systems, Man and Cybernetics. Part A, Systems and Humans, 30(3), 286–297. doi:10.1109/3468.844354 Provisions on the administration of algorithm recommendations for internet information services (2022). Retrieved from http://​www​.cac​.gov​.cn/​2022​-01/​04/​c​_1642894606364259​ .htm Raso, F., Hilligoss, H., Krishnamurthy, V., Bavitz, C. and Levin, K. (2018). Artificial intelligence & human rights: Opportunities & risks. Berkman Klein Center for Internet & Society. Retrieved from http://​nrs​.harvard​.edu/​urn​-3:​HUL​.InstRepos:​38021439 Salgado-Criado, J. and Fernandez-Aller, C. (2021, June). A wide human-rights approach to artificial intelligence regulation in europe. IEEE Technology & Society Magazine, 40, 55–65. doi:10.1109/MTS.2021.3056284 Retrieved from https://​ ieeexplore​ .ieee​ .org/​ document/​9445794 Senge, P. M. (2006). The fifth discipline (2nd edition). London: Random House, Business Books. Tambe, P., Cappelli, P. and Yakubovich, V. (2019). Artificial intelligence in human resources management: Challenges and a path forward. California Management Review, 61(4), 15–42. doi:10.1177/0008125619867910 Tribunale Ordinario di Bologna (2020). Judgment of the Tribunale Ordinario di Bologna of 27.11.2020. Available: https://​www​.bollettinoadapt​.it/​wp​-content/​uploads/​2021/​01/​ Ordinanza​-Bologna​.pdf. Troncoso Reigada, A. (2010). Las redes sociales y la APDCM. Datospersonales.org: La revista de la Agencia de Protección de Datos de la Comunidad de Madrid, (43) Retrieved from http://​dialnet​.unirioja​.es/​servlet/​oaiart​?codigo​=​3141865

6. Algorithmic management from a ‘fault line’ to a frontline opportunity for trade unions through organizational learning pragmatist take Pierrette Howayeck

INTRODUCTION The world is currently undergoing a series of shocks inducing radical changes affecting people’s lives in every aspect and specifically their daily work experience (Ferreras et al., 2020). However, what stands out from these shocks is an awareness that, on the one hand, institutions are falling behind, not being able to address the rise of new challenges on the labour market and, on the other hand, that policymakers demonstrate slow reactivity and a lack of creativity to remedy the situation with an adequate institutional framework (Lévesque et al., 2020; Ferreras et al., 2020). In their seminal paper, Murray et al. (2020) named this reality by labelling it ‘fault lines’ while rolling out a list of its seven sources with ‘technological disruption’ being – expectedly – on top. Accordingly, algorithmic management (AM) seems to fit the bill; especially that described in a recent report published by the EU-ILO Project; Baiocco et al. (2022) pointed out its ‘detrimental effects’ on employment relations and workplace bargaining due to the power imbalance between workers and management when these tools are deployed. Interestingly, Holubová (2022) reports a series of alerting numbers, ranging from 80% of workers being aware of AM risk to only 22% of them being asked for their consent before implementation and, most importantly, a total absence of knowledge among workers of any agreement or specific measure to regulate AM. Therefore, for trade unions ‘to remain agents of social justice, it is crucial that they develop the capacity to respond meaningfully to such developments’ (Flanagan and Walker (2021). Throughout this chapter we argue that developing organisational learning could allow trade unions (TUs) to take back a meaningful role, especially when taken with a pragmatist approach where it becomes a process to step into unknown territory to ‘face mystery’ (Gherardi, 1999) instead of the acquisition of a previously gained knowledge (Brandi and Elkjaer, 2012). We will seek, across this chapter, to highlight how a pragmatist take on OL can help transform a fault line into an opportunity for TUs. This chapter is structured as follows. First, we will, highlight and frame a definition of AM and the limitations of current regulation and negotiation. Next, we will explore the OL literature and its various approaches while detailing the pragmatist 74

Algorithmic management from a ‘fault line’ to a frontline opportunity  75

take. Following this, we will highlight the importance of learning for TUs and how AM can pave the way for a series of opportunities and a repositioning for TUs through OL.

ALGORITHMIC MANAGEMENT: A FAULT LINE NOT TO NEGOTIATE OR REGULATE An abundant literature defines AM with several nuances, however, for us and for the sake of clarity, we retain Meijerink and Bondarouk (2021)’s definition, which considers AM as a ‘system of control that relies on machine-readable data and software algorithms that support and/or automate managerial decision-making about work’ (Meijerink and Bondarouk, 2021, p. 3). There are two key elements to pinpoint in the retained definition. The first it that it does not hold a deterministic vision of AM as solely a tool of automation, but as also a tool to support decision-making, reflecting the current reality of many implementations. The second key element with this definition is the notion of control, which was earlier developed by Kellogg et al. (2020) who identified six main mechanisms labelled the ‘6 Rs’,1 allowing employers to control workers through an algorithm. It is essential to note that control can generate resistance if no institutional boundaries are set, creating, therefore, what Kellogg et al. (2020) described as ‘algoactivisim’, englobing many collective and individual methods to resist algorithmic control. Ironically this aversion is far from being new in managerial practices, with Northcott et al. (1985) already pointing out that when new technologies are considered by companies, control optimization is set to be a priority to improve product quality and cost reduction. However, on the less bright side of the coin, ‘labour processes, human resources and industrial relations are invariably of secondary importance and are often considered only in the implementation stage, rather than initially when making high-level corporate decisions about technological innovation’ (Lansbury and Bamber, 2013). The latter statement is more powerful and complex in the context of the introduction of AM than what was the case in the past with the introduction of new machines and automation leading to job losses. The implementation of AM necessitates a comprehensive reconfiguration within the organization, resulting in a shifting power dynamic between managers, workers and algorithms. This transition raises concerns regarding organizational accountability, as well as the diverse range of options available to managers in terms of AM objectives (such as performance reviews, incentive allocation and departure alerts). Furthermore, the identification of new roles and competencies in this context appears to be incomplete (Jarrahi et al., 2021). With all this in mind, we can witness how AM is truly a ‘fault line’, which, as defined by Murray et al. (2020), ‘disrupt[s] traditional modes of work regulation, compelling actors in the world of work to come up with strategies as best they can, but also opening up spaces for experimentation in the major arenas for the regulation of work and employment’. The ‘race to AI’ has brought forth a ‘race to AI regulation’ (Smuha, 2021) to ‘establish a relevant and protective AI governance framework’ (Ponce Del Castillo, 2020) able to set clear bound-

76  Research handbook on human resource management and disruptive technologies

aries for the data and algorithm-powered technologies. Central to any discussion on AI regulation comes the pioneering initiatives set by the EU. Nevertheless, one of the main limitations of these initiatives, as extensively explained by many labour policy experts (Ponce Del Castillo, Aloisi, De Stefano, Taes, etc), is their broad scope and/ or the absence of any straightforward clauses regarding how AM tools should be permitted to protect workers from ‘algorithmic bosses’ (De Stefano, 2020). In the case of the General Data Protection Regulation (GDPR), whose main aim is to address the ‘imbalances between those who have the ability to collect data and the data subjects’ (Ponce Del Castillo, 2020), while aware that, in order to perform as AI, systems are data-hungry. However, and as depicted by De Stefano (2020), ‘The GDPR, is no panacea against the excesses of management-by-algorithm and the use of AI at the workplace’; confirming this statement was article 88, which ‘was supposed to be a standalone piece of legislation but the EU Commission eventually decided against this idea’ (Ponce Del Castillo, 2020). In fact, article 88 of the GDPR recommends without forcing the EU Member States to have the possibility to introduce, by law or by collective agreements, ‘specific rules to ensure the protection of the rights and freedoms in respect of the processing of employees’ personal data in the employment context’ (De Stefano, 2020). Moreover, the European Commission published (in April 2021) the AI Act draft with a risk-based approach grounded on the pyramid of criticality. In other words, the AI Act stipulates that, to regulate in an effective way one solution would be to scale AI between low and high risk, in order to ensure proportionate intervention while keeping in mind the necessity of a combination of two cumulative criteria of high-risk AI: the sector of activity with which this application is concerned and the concrete use of AI in this sector (Blin-Franchomme and Jazottes, 2021). This AI Act engenders within it a series of policy issues, as identified by Ponce Del Castillo (2021), ranging from failing to address the use of AI in the world of work, to authorizing automatically without any evaluation AI applications categorized as low risk, and finally and most importantly, granting permission to use high-risk subjects if they comply with a set of conformity assessments, putting therefore tech providers’ interests over EU citizens and workers’ rights. Therefore, in the absence of a specific regulation in law (Ajunwa et al., 2017; Todolí-Signes, 2019) the governance of AI could not be assigned to a handful of actors, especially given that the European ambition is to impose themselves as the global digital leader – an ambition that doesn’t put at stake its ‘fundamental rights, social dialogue and tripartite participation’ (Ponce Del Castillo, 2020). AM is therefore a complex object of regulation, while actual regulation falls behind; in setting clear boundaries it needs to also maintain a certain degree of flexibility to ensure high adaptability to the moving ground of developing technology (De Stefano, 2018; Todolí-Signes, 2019). Thereby, the TU role and, precisely, collective bargaining is arguably still the most effective tool to provide safeguards against the rapid technological developments in AM (De Stefano, 2020). It is essential to note that ‘negotiating the algorithm’ (De Stefano, 2018) is a crucial objective of social dialogue, which does not come without a series of major challenges (Bernier, 2021; Hennebert and Bourguignon,

Algorithmic management from a ‘fault line’ to a frontline opportunity  77

2021). These challenges start with the fact that TUs should demonstrate the ability to identify and plan adequate responses to the risks of AM (De Stefano and Taes, 2021); however, we know that this does not seem to be the case in many situations, as indicated by Amauger-Lattes (2021) and through our extensive exploratory study. In fact, AI brings ‘many anxieties and misunderstandings [that] paralyze discussions on digital transformation in companies’ (Jolly and Naboulet, 2017) as ‘apprehending objects as unstable and evolving as algorithmic systems is not easy for anyone’ (Leroy, 2020). Therefore, addressing the challenges of AI starts with employment and working conditions but can be extended to issues involving fundamental rights where the obligations of dialogue are limited (Amauger-Lattes, 2021). In this respect, transnational collective bargaining and reports on transnational agreements may potentially play a key role to ease these challenges (Seifert, 2018) and offer a more methodological approach, as is the case with the European Social Partners Framework Agreement on Digitalization. Even though this agreement is broad in scope and not limited to AI, it serves as a guide to social partners engaged in these negotiations, drawing their attention to potential pitfalls and emphasizing rights to be guaranteed (Amauger-Lattes, 2021). To wrap things up, we have gone briefly through the major and non-constraining regulations that have left more space to autoregulation by companies and, in other cases, disregarded workers’ protections. Moreover, we revisited the difficulty of having a solid negotiation around AM, with the proof that the European Social Partners Framework Agreement on Digitalization is far from giving any concrete assurance. Therefore, we suggest that TUs need to act differently to grasp the technicity of the subject and impose themselves as experts, having in their hands practical solutions to advance negotiation and, ultimately, regulate. We believe that, in order to achieve this, TUs need to learn the ins and outs of the subject, accumulating knowledge from all the persons engaged in any AM implementation, including solutions vendors, consultants, decision-makers inside a company, impacted workers, etc. This dynamic of learning could only be achieved, as we will see below, through a pragmatist take on organizational learning, where TUs will act differently to have a strategic impact on the future of work.

ORGANIZATIONAL LEARNING OL has been dense in recent years with literature reviews that, instead of creating harmony in the field, have spawned divergent approaches and created a never-ending debate (Koenig, 2015; Ingham, 2015). The genesis of organizational learning began in the field of management, specifically through theories of organizational behaviour with the pioneering works of March and Simon (1958) and Cyert and March (1963). Early literature emerged from the need of organizations to adapt to constant changes, therefore providing a more prescriptive managerial technique. Three decades later, with the publication of the book Learning Organization, Senge tried to concretize the way toward OL creation, enforcing more this managerial view of the field (Brandi and Elkjaer, 2012; Senge, 1990). Interestingly, surfing this literature is far from

78  Research handbook on human resource management and disruptive technologies

being an easy mission, with many proliferations, and literature reviews that struggle to align and harmonize around the same seminal papers as foundations of the field. It wasn’t until 2003 when Elkjaer brought light into the abyss, when, in introducing a new approach to the field, she offered a new way to understand and categorize the abundance of approaches in OL. In order to categorize the diverse OL approaches, Elkjaer used Sfard (1998) to present an understanding of learning with two metaphors: ‘knowledge acquisition’ and ‘knowledge participation’ (Elkjaer, 2003). According to Elkjaer (2003), ‘Metaphors can provide an opportunity for a new vision of phenomena. Thus, these two metaphorical understandings of learning allowed me to see the field of OL in a different light’. To set the ground for Elkjaer’s third metaphor, or ‘the third way’ as she also designates it, we will first go through the two previously evoked metaphors, paving the way to see better how the third way would be the best fit in our case. An Overview of the ‘First and Second Way’ of Approaching Organizational Learning Our primary reflex toward understanding learning was for a long time associated with teaching and education, or simply to knowledge acquisition, a fact that influenced early OL scholars’ and their contributions to the literature. As highlighted by Elkjaer (2003; 2004), the acquisition metaphor was the ‘prevalent basis for the understanding of learning in the field of OL’, with Argyris and Schon (1996) and March and Simon (1958) considering that achieving OL is through ‘individuals’ acquisition of information and knowledge, analytical and communicative skills’ (Elkjaer, 2004). Consequently, and following much criticism, a question has arisen (Cook and Yannow, 1993; Elkjaer, 2003) on how concretely a transfer occurs between what an individual has acquired through the learning process and the organization, therefore resurfacing the problem of separating individual and organization. On the other hand, Argyris and Schon considered that ‘OL occurs when individuals within an organization experience a problematic situation and inquire into it on the organization’s behalf’ (Argyris and Schon, 1996) – a tentative way of solving the individual and organization distortion, which only found a new way to confirm it. Moreover, OL literature was not explicit and concrete around this notion of acting on behalf of an organization in order to validate the ‘understanding of individuals and organization as two entities that can be understood partially in a separate way’ (Elkjaer, 2004). Through this acquisition metaphor we were more in a cognitive process where learning is ‘taking place in the minds of the individuals’, knowing that, as emphasized by Elkjaer (2004), ‘learning is a practical rather than a cognitive process’. From this point we go back to Sfard’s metaphors where, as explained by Elkjaer (2003), it was constructed through its direction of educational research, covering learning in educational institutions following a teaching experience where individuals and groups took centre stage. Elkjaer (2003) transcended Sfard’s focus on groups and participation metaphor to highlight that ‘there is an element of “community” and collective work practice in an

Algorithmic management from a ‘fault line’ to a frontline opportunity  79

organization, which employees are part of, and which is the frame of reference for OL’ (Elkjaer, 2003), echoing, therefore, the work of Lave and Wenger (1991). A key element is that Elkjaer’s inspiration of the participation metaphor in OL literature was principally due to Lave and Wenger’s works, which considered that learning occurs while participating in what they labelled in their influential book as ‘communities of practice’ (CoP), making it then a ‘legitimate peripheral participation in CoP’. It is important to highlight that organizational literature adopting the participation metaphor, having the patterns of collective participation as a ‘unit of learning’ instead of individuals (Elkjaer, 2004), was keen on putting the accent on the fact that learning is not limited to the cognitive process (acquisition metaphor) (Gherardi et al., 1998) nor is it an activity that surges occasionally, but rather a daily one embedded in organization of life and work practices (Nicolini and Meznar, 1995). CoPs are understood as groups of people who are interconnected by means of informal relationships interacting regularly (Sena and Shani, 1999) and sharing the same practices because they care about the same real-life problems or hot topics (Wenger et al., 2002). Through these CoPs, a potential increase in the level of knowledge transfer is believed to take place (Dewhurst and Cegarra-Navarro, 2004) as they provide an ‘informal learning environment in which novices and experienced members of the community may interact with each other, share their experiences of being in a particular profession, and learn from each other’ (Hara and Schwen, 2006). Hence, one key element in CoPs is that they are not strictly limited to members in the organizational borders, especially in times of crisis, when fault line surges push us not to apprehend them as inducers of change but rather as an opportunity to invite new protagonists with new strategies, thus shifting things (Dewhurst and Cegarra-Navarro, 2004; Voss and Sherman, 2000). According to Slee (2014), driving change requires ‘innovators’, individuals with the ‘knowledge, vision and sense of urgency’ required to overcome organizational obstacles. Having this in mind makes it clear that creating intellectually heterogeneous groups would enhance the ability to overcome these obstacles, especially considering that, with the ‘group think’ model, they are more prone to be innovative than homogeneous ones (Leonard and Sensiper 1998, p. 118). To sum up, as acknowledged by Brown and Duguid (1991), organizational members need to become knowledgeable to face challenges, and it is through considering OL through the lens of a social process of participation, construction of CoPs and memberships that this is achieved. Having overviewed these two ways of approaching OL, we sum up with some of Elkjaer’s key observations that will pave the way to the third way. OL in its participation metaphor seems much suited for organizational contexts, especially in that it goes beyond the individual scope, reaching work practice and organizations while emphasizing that teaching and individual knowledge were not the destination. However, ‘...the “how” and the “what” of learning seem to disappear within the broader concept of “learning as participation”. In other words, how is learning taking place and what is learned by way of participating in communities of practice?’ (Elkjaer, 2004).

80  Research handbook on human resource management and disruptive technologies

The ‘Third Way’ of OL: ‘The Pragmatist Take’ From what we previously emphasized as the backlashes against the second way of OL – the ‘how’ and ‘what’ questions – an answer was found by Elkjaer in the work of the American pragmatist and educationalist, John Dewey. Elkjaer acknowledges that Dewey’s notions on inquiry and experience allow a better understanding of these two questions, without being able to frame ‘organizational dynamics in which learning is situated’, also considered a crucial point (Elkjaer, 2004; 2007). In order to remedy this deficiency, Elkjaer took additionally the concept of a social worlds perspective from the pragmatic theory of organizations as a point of departure where commitment to action and interaction are emphasized, especially in that organizations are understood as systems in the ‘first way’ and as CoP in the ‘second way’. This combination allowed the development of the ‘third way’ of OL, in which learning is understood as ‘emerging’ or ‘becoming’; therefore, no privileging was given to one over another of the two aforementioned understandings (Sprogoe and Elkjaer, 2010). Consequently, pragmatism offers a set of ‘conceptual tools’ (experience and inquiry as the ‘what’ and ‘how’ of learning) to create a new organizational learning approach that ‘synthesize[s] the “second way”, with its understanding of learning as participation in CoP with elements of the “first way” related to learning as the acquisition of skills and knowledge’ (Elkjaer, 2004). We propose that an adequate exploration of Elkjaer’s ‘third way’ would best be done following a quick concept revisiting of what shaped it from inquiry, experience and social worlds. Furthermore, this concept revisiting will also pave the way to transcend it at a later stage as the basis of why to approach the AM and TU dilemma through the third way. According to Dewey (1916 [1980]), ‘thinking is a process of inquiry, of looking into things, of investigating. Acquiring is always secondary, and instrumental to the act of inquiry. It is seeking, a quest, for something that is not at hand.’ This extract of Dewey’s work offers the essence of learning in the pragmatist understanding or, as expressed by Elkjaer (2007), ‘to create meaning in and with uncertain situations, which involves first to define it as a problem and doing that needs positioning (perspective) as well as reasoning’. Subsequently, inquiry is defined as ‘the controlled or directed transformation of an indeterminate situation into one that is so determinate in its constituent distinctions and relations as to convert the elements of the original situation into a unified whole’ (Dewey 1936 [1986]). Transferring Dewey’s inquiry into understanding onto OL paves the way for apprehending it as a process allowing the creation and implementation of knowledge through the transformation of uncertain situations facing organizations into a more stable one, resulting potentially in a change in organizational habitual actions (Elkjaer, 2017). Uncertain situations are, consequently, the trigger at the beginning of any learning journey; it is nor individuals or organizations that are responsible for it, as claimed, respectively, in the first and second ways (Elkjaer and Simpson, 2011). It is also important to recognize that any situation is a combination of individuals as well as organizational contexts though the ‘focal point is the problematic situation’ (Sprogoe and Elkjaer, 2010).

Algorithmic management from a ‘fault line’ to a frontline opportunity  81

Having all this in mind, we come back to the legitimate previously raised question when exposing the first two ways of learning and how organization and individual interact. Reviewing the work of Elkjaer and Simpson (2011), a clear answer can be given: In our pragmatist understanding, the relationship between individual and organization is not guided by individuals’ choices but by the transactional interplay between the two. The meanings attached to individuals and organizations are, therefore, highly interdependent, and continuously evolving. The transactional approach draws together subjects (individuals), objects (knowledge) and situations into a mutually constituting, dynamic whole. To see OL as fundamentally transactional is to focus on the interplay between selves and situations rather them as discrete levels in the social system. If these levels are treated separately, then we are left with the intractable problem of having to glue them together.

Moving now to define experience in Dewey’s sense, where the transactional approach is echoed and ‘subtleties of inquiry concept’ are held (Elkjaer, 2021), a definition that diverges considerably from our traditional concept of experience. Interestingly Buch and Elkjaer (2020) captured the essence of this definition from Dewey’s ‘The Need for a Recovery of Philosophy’ (1916 [1980]) via the following quote: (...) experiencing means living; and that living goes on in and because of an environing medium, not in a vacuum. (...) Experience is primarily a process of undergoing: a process of standing something; of suffering and passion, of affection, in the literal sense of these words. (...) Experience is no slipping along in a path fixed by inner consciousness. (...) Since we live forward; since we live in a world where changes are going on whose issues means our weal or woe; since every act of ours modifies these changes and hence is fraught with promise, or charged with hostile energies – what should experience be but a future implicated in the present! (Dewey, 1916 [1980]: pp. 7–9 [Dewey’s punctuation]).

Subsequently, and according to Illeris (2009) one can depict the following divergences between Dewey’s concept of experience and a traditional one: knowledge as a subset of experience rather than experience as knowledge; experience as both subjective and objective rather than solely subjective; experience as future-oriented instead of past; experience as united experiences not as isolated ones; and finally, experience as encompassing theories and concepts as a foundation for knowledge rather than action. Having clarified the pragmatist understanding of experience, it is now essential to see how this crystallizes our previous inquiry understanding. For this purpose, we move on now toward exploring experience and inquiry in relation specifically to Brandi and Elkjaer (2018)’s complete summary: It is in experience that difficulties arise, and it is with experience that problems are resolved by inquiry. Inquiry is an experimental method by which new experience may be had not only through action but also by using ideas and concepts, hypotheses, and theories as ‘tools to think with’ in a playful and instrumental way. Inquiry is concerned with consequences and pragmatism views persons as future-oriented rather than oriented towards the past.

82  Research handbook on human resource management and disruptive technologies

To sum up, difficulties are revealed through transaction while experiencing and it is through inquiry that one gains the experience to solve problems (Elkjaer, 2018). Turning now to organizations, they are shaped with individuals through the relationship they have, known as ‘transactional relationships’ (Elkjaer, 2004; Altman and Rogoff, 1987; Dewey and Bentley, 1949 [1991]). According to Elkjaer (2004) the ‘transactional’ concept can also be described as ‘relationist’, which makes it possible to see ‘the social world as dynamic relations that unfold’ (Elkjaer, 2004; Emirbayer, 1997; Nicolini et al., 2003). To wrap up Elkjaer’s ‘third way’ of OL, we will go through defining social worlds, which, according to Clarke (1991; quoted in Strauss, 1993), are ‘groups with shared commitments to certain activities, sharing resources of many kinds to achieve their goals, and building shared ideologies about how to go about their business’. Taking the works of Clarke (1991) and Strauss (1978; 1993) as references, social worlds are summarized by Huysman and Elkjaer (2006) as follows: ‘works is understood as “coordinated collective actions” and organisations are understood as “arenas” of “social worlds” created and maintained by commitment to organisational activities’. Moreover, Huysman and Elkjaer (2006) emphasized the following points on social worlds, bringing a clearer understanding of their role in the OL ‘third way’: ● Social worlds limits are effective communication considered as interactive units not constrained by membership or geography, a ‘universe of regularized mutual response, communication or discourse’ (Shibutani, 1955). ● Social worlds have the influence to ‘inform you what knowledge is important and what knowledge is not’ (Huysman and Elkjaer, 2006). ● Social worlds and CoP both allow one to witness collective learning however the latter put the accent on harmony when the first makes visible how ‘various issues are debated, negotiated, fought out, forced, and manipulated by representatives’ (Strauss, 1978). A final word on OL as a ‘third way’ is that it is not a replication of learning in the sense of the ‘second way’, with Lave and Wenger’s CoP, even though it shares a similar understanding; however, pragmatism adds two important elements: ‘an understanding of what triggers learning (the uncertain situation) as well as what produces community or solidarity at work (commitment to actions, activities and values)’ (Sprogoe and Elkjaer, 2010).

TRADE UNIONS AND LEARNING: A STRATEGIC NEED IN A CHANGING WORLD Trade Unions and Learning the All-Time Constant One of TUs’ key roles is to be mediators – as noted by Müller-Jentsch (2018), who evokes the works of Bergman et al. (1975/1979) – the fulfilling of which depends

Algorithmic management from a ‘fault line’ to a frontline opportunity  83

on intermediating the interests of capital and those of the workforce. Consequently, they were labelled ‘intermediary organizations’, passing therefore from a ‘fighting’ to an ‘institutional’ status (Briefs, 1952), which contributed to making them ‘strong and influential for three reasons: on account of their innate power, on account of their legal recognition and on account of their acceptance among employers’ (Briefs, 1952, cited by Müller-Jentsch, 2018). Accordingly, through the 20th century, TUs were envisioned as national actors, collective bargainers and even ‘social partners’, aiming ‘to influence the macroeconomic and social policies of national governments’ (Hyman, 2007). This trend, however, was hindered by a series of complex and contested reasons (Walker et al., 2007; Lucio, 2006), of which we emphasize globalization, which weakened the national level of employment regulation, and capitalism, which urges constant restructuring and transformation to survive (Hyman, 2007). The aftermath of all this left TUs unable to maintain their role as actors of collective representation (Dufour and Hege, 2010) (especially with diminishing membership), instead assuming the role of ‘fire-fighters, reacting desperately to challenges to the established “industrial legality”’ (Hyman, 2007). To counteract this deplorable situation, many discussions pointed out the urgent need for TU revitalization, modernization and renewal (Hälker and Vellay, 2006). All of this led various scholars to pinpoint the importance of a renewal strategy by suggesting models allowing TUs to curb this trend through effective response to external pressure while maintaining an introspective vision of their internal environment. We will go briefly through three of these models. First, Hyman (1997)’s ‘organizational capacity model’, which ‘can be understood as the ability to assess opportunities for intervention; to anticipate, rather than merely react to, changing circumstances; to frame coherent policies; and to implement these effectively’ (Hyman, 2007). The key components of the ‘organizational capability’ are a structure considered as the ability to condense into one set of priorities different TU perspectives, intelligence or the ability to ‘possess specialist expertise in research, education and information-gathering, and the means to disseminate knowledge throughout the organization’ (Hyman, 2007); a strategy that ensures a coherent link between knowledge and action through assessment, analysis and planning of situations, alternatives and objectives, respectively; finally efficacy is the achievability of objectives fixed. Second, Frost (2000)’s model identifying four capabilities ‘accessing information, educating and mobilizing the membership, accessing management decision making at multiple points, and balancing cooperation and conflict’ (Frost, 2002). This model was complemented by Frost herself with five national unions characteristics to sustain these capabilities, described as follows: ‘the characteristics are the breadth of the national union’s representational coverage; the extent of its education and training focus on new workplace issues; the resources it devotes to research on the implications of workplace practices; the presence of multiple communication channels; and its structuring of local union representation’ (Frost, 2002). The third and last model we will be covering is Lévesque and Murray’s ‘power resources’ model, which gave a ‘fresh approach to understanding of TU strategy’ (Slee, 2014). Four

84  Research handbook on human resource management and disruptive technologies

key power resources form the pillars of this framework: ‘internal solidarity; network embeddedness; narrative resources that frame understandings and union actions; and infrastructural resources (material, human, processes, policies and programs)’ (Lévesque and Murray, 2010). However, the authors consider these resources as insufficient, and TU must complement them with four strategic capabilities: ‘intermediating between contending interests to foster collaborative action and to activate networks; framing; articulating actions over time and space; and learning’ (Lévesque and Murray, 2010). Reviewing these models allows us to echo Carter’s (2000) work as summarized by Slee (2014): ‘Unions are neither passive recipients of the external environment nor can they hope to act as independent agents free of the constraints of their external environment’. With Carter’s reflection, we can grasp that learning is a recurrent necessity in all three models reviewed above, which resonates with Smith (2021) (who approached TU OL through CoP): ‘It is somewhat surprising, then, that given the concept’s telos,2 and the recognition of the importance of OL as a perspective of union innovation and renewal, that these ideas have not been coalesced, perhaps due to the apolitical; nature of the “turn to organizing” path the renewal debate tool’ (Simms and Holgate, 2010). Trade Unions and Organizational Learning TU learning has been the subject of several studies aiming to depict its relationship with union revitalization studies (Mustchin, 2012; Wallis et al., 2005); however, the main focus was on ‘formal pedagogical practices’ (Smith, 2021) without any further advancement or contribution to the OL approach for TU. Through our review of OL literature, we therefore reiterate Hyman (2007)’s conclusion that ‘OL has had little impact on TU analysis, but its relevance is not hard to perceive’ with a literature ‘overwhelmingly managerial in its approach’ (Huzzard, 2000). Even though, OL and TU literature is scarce, we have a handful of industrial relation scholars’ works that allows us to depict how it was approached overtime. Zoll (2003) employed Bateson (2000)’s model consisting of making a nuance between two learning orders. The first order relies on individuals using trial and error to solve problems until acquiring the needed experience to directly apply adequate solutions. The second order is more about learning how to learn and finding solutions quickly. However, as pointed out by Hyman (2007): Zoll goes [on] to suggest the need for third-order learning, involving the critical scrutiny and redefinition of unions’ existing learning strategies and structures and more fundamentally of their existing understanding of what is to be a TU.

Given the aforementioned considerations, it becomes evident that Zoll’s approach is more suitable in comparison to the acquisition approach that we previously emphasized. Notably, the Bateson model served as the foundation for the analysis conducted by Argyris and Schon (1974), albeit with a minor distinction. Argyris and

Algorithmic management from a ‘fault line’ to a frontline opportunity  85

Schon focused their examination on business organizations, where individuals act on behalf of the organization, while Bateson’s model centred on animal and human behaviour. However, it is worth noting that the statement in question is incomplete and appears to imply that Bateson’s model preceded that of Argyris and Schon (1974), which is inaccurate as Bateson’s model predates it by the year 2000 (Hyman, 2007). Other ways of approaching TU organization were proposed by Cooper (2006), Ball (2003) and later, Smith (2021), conceptualizing them as CoP where ‘the TU may therefore be seen as a community of practice ... where the process of participation in routine union activities ... induct workers into their trade unionist roles and identities’ (Cooper, 2006; cited by Smith, 2021). As Smith (2021) also emphasizes, ‘CoP learning is a collective endeavour, and community participants are knowledge generators, all with something to contribute: an egalitarian view befitting TU doctrine’. Even though CoP offered a ‘radical critique of cognitivist theories of learning’ (Handley et al., 2006, p. 641), Huzzard shed light on it being largely unitarist (Huzzard, 2001). Hence, Huzzard and Östergren (2002) paved the way in their work to a more pragmatist approach of learning that could be summarized by the following passage: It seems more fruitful to conceptualize learning in a way that recognizes diversity and conflict, thereby shifting the emphasis onto learning through exploration and away from learning through exploitation. In our view a useful approach here is to extend throughout the organization the concept of learning ‘arenas’ introduced by Burgoyne and Jackson in the context of management learning (Burgoyne and Jackson, 1997). The key is to design such arenas to be participative and based on the idea of democratic dialogue (Gustavsen, 1992). Under democratic dialogue, consensus is not an ex ante prerequisite of learning and is not predefined or targeted by top management. It may, however, be a serendipitous outcome of dialogue (Huzzard and Östergren, 2002).

Algorithmic Management as Opportunity for Trade Unions: Unlock it Via the ‘Third Way’! From Huzzard and Östergren (2002)’s call to explore learning through the lens of exploration not of exploitation we witness that the premises of critics for the first way and second way of OL could be found in the works of industrial relations scholars slightly before being stabilized by Elkjaer and the pragmatist approach. Hence, Huzzard and Östergren (2002) highlighted that diversity makes approaching OL with a unitarist take very problematic, especially if we consider learning a dialectical process where differences can nourish reflection and learning. In a later study, Breindenjso and Huzzard (2006) emphasized the role of reflection and that nourishing it through a genuine dialogue ‘cannot occur in traditional industrial arenas’, therefore, they suggested the establishment of what they entitled ‘learning spaces’. A concept resonating with social worlds finalities; however, with a slight difference that it is: (…) designed and moderated through unions expertise where dialogue and reflection are central, can thus be platforms for putting critical questions to firms in ways that can have developmental potential. The aim for unions here is not defend or advance the sectional

86  Research handbook on human resource management and disruptive technologies

interests of employees. Such activity will continue as core union work but will be undertaken elsewhere. Learning spaces might conceivably involve joint reflection on decision to outsource, the trajectory of technology, job redesign, customer and supplier relations and various risk analysis. On the other hand, it may well be the case that some outcomes of such reflection may need to be bargained over thereby transferred to other arenas’ (Breindenjso and Huzzard, 2006).

Having this industrial relation approach of learning makes Elkjaer’s ‘third way’ befit the needs of TU faced by fault lines. Therefore, we can see that AM that fulfils adequately the ‘fault line’s’ role can be a subject of learning and more an opportunity for TU through which they can create ‘social worlds’ or ‘learning spaces’, allowing, through inquiry, shared experience to create better understanding in a neutral situation. In fact, social worlds in our case can be described as rich spaces where heterogeneous actors (ranging from consultants, representatives of software companies, employees that experienced the implementation, HR and tech employee responsible of implementation and different trade union representatives) of AM implementation can be found around a roundtable initiated and moderated by a TU. Therefore, TUs, through this initiative, will get to playback their mediator role and be active social actors. Moreover, it will allow them to grasp a well-rounded knowledge of AM through concrete example identifying its pitfalls. Hence these spaces may result in some publications, as is the case in France, with some TUs intuitively launching these initiatives that flourished by first giving some concrete tools for trade unionists (charts, best practices report) to help raise their awareness and their negotiation skills when a tool is implemented, flourishing now on several levels. Second, it allowed TUs to have an expertise in public hearings or, while commenting, draft laws nationally and internationally. Third, it allowed TUs to achieve more cooperation with other European TUs, thus expanding the learning spaces, especially recognizing that AM is a universal problem and, therefore, better negotiation and afterward regulation can be undertaken. Finally, it also allowed some TUs to review their internal structure through dedicating a position for AM assuming its strategic impact in the long term if not treated well. To wrap up, we can see that AM offers not only an opportunity, ‘allowing TU to face a hostile external environment threatening them but also “providing the platform for renewal”’ (Kelly, 1998, quoted in Fairbrother et al., 2007).

CONCLUSION AND RECOMMENDATIONS FOR ACTION According to Slee (2014), negative shocks allow trade unions to break with their past strategy offering a more ‘fertile environment’ where new choices are made to face adequately the situation. Through this chapter, we highlighted how AM as a ‘fault line’ has hardly been approached with regulation and negotiation as a first step due to its novelty, complexity, and the proliferation of actors and actions it engenders. Holubová (2022) highlights that, regarding AM, workers expect TUs to tackle three main priorities: increasing workers’ awareness, securing ethical and socially respon-

Algorithmic management from a ‘fault line’ to a frontline opportunity  87

sible development of AM and ensuring strong collective bargaining. Therefore, TUs are to act differently, especially given that not one of the abovementioned priorities is achievable without having the right knowledge. Consequently to adapt to change and develop skills that make it possible to face new challenges while catching new opportunities, one should acknowledge the role of ‘creation and dissemination of new knowledge’ (Slee, 2014) and therefore of OL in achieving all this. On a general basis, we went through three approaches of OL; the first considers a cognitive process, where individuals act as agents learning on behalf of the organization, the second takes a participatory approach through the concepts of OL, and the third is the transaction process, which we found to be adequate for TUs dealing with AM with its three pragmatist pillars as identified in Elkjaer’s works, inquiry experience and, more importantly, social worlds. Considerably more work needs to be done by scholars to explore empirically this pragmatist approach of learning in TUs, knowing that some has already taken place in France with European financing and cooperation. This approach would be a fruitful area for further work on the role that other power resources and capabilities (Lévesque and Murray, 2010) could play in facilitating the construction of such spaces. Hence, we highly encourage TUs to consider constructing such spaces and including actors with divergent interests, as this makes it possible to shape knowledge in a less conflicting context. However, future research studies will aim to segregate some best practices and lessons to help them concretely launch this ahead of disruption while counteracting some pitfalls already detected from previous experience. Further studies will also highlight how such spaces can allow the unity of several TUs who could ideologically be different, condensing inside one project solutions for a better future for workers while collaborating with management, consultants and solutions vendors, etc.

NOTES 1. ‘Employers can use algorithms help to direct workers by restricting and recommending, evaluate workers by recording and rating, and discipline workers by replacing and rewarding’ (Kellogg et al., 2020). 2. Smith (2021) was referring to the CoP concept: ‘In COP learning is a collective endeavour, and community participants are knowledge generators, all with something to contribute: an egalitarian view befitting trade union doctrine.’

REFERENCES Ajunwa, I., Crawford, K., and Schultz, J. (2017). Limitless worker surveillance. California Law Review, 105, 735. Altman, I. and Rogoff, B. (1987). World views in psychology: trait, interactional, organismic, and transactional perspectives. In D. Stokols and I. Altman (eds), Handbook of Environmental Psychology (Vol. 1), John Wiley & Sons, pp. 7–40. Amauger-Lattes, M. C. (2021). Le dialogue social, outil de régulation de l’intelligence artificielle dans l’entreprise. Droit social, (2), 146–51.

88  Research handbook on human resource management and disruptive technologies

Argyris, C. and Schon, D. A. (1974). Theory in Practice: Increasing Professional Effectiveness. Jossey-Bass. Argyris, C. and Schön, D. A. (1996). Organizational Learning II: Theory, Method, and Practice. Addison-Wesley (vol. 1). Baiocco, S., Fernández-Macías, E., Rani, U., and Pesole, A. (2022). The algorithmic management of work and its implications in different contexts (No. 2022/02). JRC Working Papers Series on Labour, Education and Technology. Ball, M. (2003). Considering trade union education as a community of practice. International Journal of Lifelong Education, 22(3), 297–310. Bateson, G. (2000). Steps to an Ecology of Mind: Collected Essays in Anthropology, Psychiatry, Evolution, and Epistemology. University of Chicago Press. Bergmann, J., Jacobi, O. and Müller-Jentsch, W. (1975/1979). Gewerkschaften in der Bundesrepublik, Band 1: Gewerkschaftliche Lohnpolitik zwischen Mitgliederinteressen und ökonomischen Systemzwängen, Auflage 1975, 3. Auflage 1979, Frankfurt am Main. Bernier, J. (2021). L’intelligence artificielle et les mondes du travail. Perspectives sociojuridiques et enjeux éthiques. Presses de l’Université Laval. Blin-Franchomme, M. P. and Jazottes, G. (2021). Le défi d’une IA inclusive et responsable. Droit social, (2), 100–5. Brandi, U. and Elkjaer, B. (2012). Organizational learning viewed from a social learning perspective. In M. Easterby-Smith and M. A. Lyles (eds), Handbook of Organizational Learning and Knowledge Management, John Wiley & Sons, pp. 21–41. Brandi, U. and Elkjaer, B. (2018). Subtleties of knowledge sharing – results from a case study within management consultancy. Knowledge and Process Management, 26(3), 185–93. Breidensjö, M. and Huzzard, T. (2006). Reflecting on workplace change: A trade union perspective. In Productive Reflection at Work, Routledge. (pp. 146–57). Briefs, G. (1952). Zwischen Kapitalismus und Syndikalismus. Die Gewerkschaften am Scheideweg. Francke. Brown, J. S. and Duguid, P. (1991). Organizational learning and communities-of-practice: toward a unified view of working, learning, and innovation. Organization Science, 2(1), 40–57. Buch, A. and Elkjaer, B. (2020). Pragmatism and practice theory: convergences or collisions. Caderno de Administração, 27(2), 1–17. Burgoyne, J. and Jackson, B. (1997). The arena thesis: Management development as a pluralistic meeting point. Management learning: Integrating perspectives in theory and practice, 54–70. Carter, B. (2000). Adoption of the organising model in British trade unions: some evidence from manufacturing, science and finance (MSF). Work, Employment and Society, 14(1), 117–36. Clarke, A. E. (1991). Social worlds/arenas theory as organizational theory. In D. R. Maines (ed.), Social Organization and Social Process. Essays in Honor of Anselm Strauss. Aldine de Gruyter, pp. 119–58. Cook, S. N. and Yanow, D. (1993). Culture and organizational learning. Journal of Management Inquiry, 20(4), 362–79. Cooper, L. (2006). The trade union as a learning organisation? A case study of informal learning in a collective, social-action organisational context. Journal of Education, 39(1), 27–46. Cyert, R. M. and March, J. G. (1963). A behavioral theory of the firm – 40 years and counting: introduction and impact. Organization Science, 2(4), 169–87. De Stefano, V. (2018). ‘Negotiating the algorithm’: automation, artificial intelligence and labour protection. ILO Working Papers 994998792302676, International Labour Organization.

Algorithmic management from a ‘fault line’ to a frontline opportunity  89

De Stefano, V. (2020). Algorithmic bosses and what to do about them: automation, artificial intelligence and labour protection. In D. Marino and M. A. Monaca (eds), Economic and Policy Implications of Artificial Intelligence, Springer, pp. 65–86. De Stefano, V. and Taes, S. (2021). Algorithmic Management and Collective Bargaining. Foresight Brief. Dewey, J. (1916 [1980]). Democracy and education: an introduction to the philosophy of education. In J. A. Boydston (ed.), Middle Works 9, Southern Illinois University Press, 155. Dewey, J. (1936 [1986]). The Collected Works of John Dewey, 1882–1953, Jo Ann Boydston (ed.), Carbondale and Edwardsville: Southern Illinois University Press, 1969–1991. Dewey, J. and Bentley, A. F. (1949 [1991]). Knowing and the known. In J. A. Boydston (ed.), Later Works 16, Southern Illinois University Press, pp. 1–294. Dewhurst, F. W. and Cegarra-Navarro, J. G. (2004). External communities of practice and relational capital. The Learning Organization, 1(4/5), 322–31. Dufour, C. and Hege, A. (2010). The legitimacy of collective actors and trade union renewal. Transfer: European Review of Labour and Research, 16(3), 351–67. Elkjaer, B. (2003). Organizational learning with a pragmatic slant. International Journal of Lifelong Education, 22(5), 481–94. Elkjaer, B. (2004). Organizational learning: the ‘third way’. Management Learning, 35(4), 419–34. Elkjaer, B. (2007). Organisational Learning: Positioning of Selves and Creation of Meaning. Learning Lap Denmark, pp. 180–94. Elkjaer, B. (2017). Organizations as real and ephemeral. Zeitschrift fuer Weiterbildungsforschung, 40(1), 53–68. Elkjaer, B. (2018). Knowledge production as organisational learning: the case of Danish Universities. The Palgrave International Handbook on Adult and Lifelong Education and Learning, 289–308. Elkjaer, B. (2021). The future is in the present: learning and organizing in the field of organizational learning. In M. Malloch, L. Cairns, K. Evans, and B. N. O’Connor (eds), The SAGE Handbook of Learning and Work, pp. 208–22. Elkjaer, B. and Simpson, B. (2011). Pragmatism: a lived and living philosophy. What can it offer to contemporary organization theory?. In H. Tsoukas and R. Chia (eds), Philosophy and Organization Theory (Research in the Sociology of Organizations, Vol. 32), Emerald Group Publishing Limited, pp. 55–84. Emirbayer, M. (1997). Manifesto for a relational sociology. American Journal of Sociology, 103(2), 281–317. Fairbrother, P., Williams, G., Barton, R., Gibellieri, E., and Tropeoli, A. (2007). Unions facing the future: questions and possibilities. Labor Studies Journal, 31(4), 31–53. Ferreras, I., MacDonald, I., Murray, G., and Pulignano, V. (2020). L’expérimentation institutionnelle au travail, pour le meilleur (ou pour le pire). Transfer: European Review of Labour and Research, 26(2), 119–25. Flanagan, F. and Walker, M. (2021). How can unions use artificial intelligence to build power? The use of AI chatbots for labour organising in the US and Australia. New Technology, Work and Employment, 36(2), 159–76. Frost, A. C. (2000). Explaining variation in workplace restructuring: the role of local union capabilities. ILR Review, 53(4), 559–78. Frost, A. (2002). Creating and sustaining local union capabilities the role of the national union. Relations industrielles/Industrial Relations, 56(2), 307–35. Gherardi, S. (1999). Learning as problem-driven or learning in the face of mystery?. Organization Studies, 20(1), 101–23. Gherardi, S., Nicolini, D., and Odella, F. (1998). Toward a social understanding of how people learn in organizations: the notion of situated curriculum. Management Learning, 29(3), 273–97.

90  Research handbook on human resource management and disruptive technologies

Gustavsen, B. (1992). Dialogue and Development. Van Gorcum. Hälker, J. and C. Vellay (2006). Union Renewal: Gewerkschaften in Veränderung, Düsseldorf. Hans Böckler Stiftung. Handley, K., Sturdy, A., Fincham, R., and Clark, T. (2006). Within and beyond communities of practice: making sense of learning through participation, identity and practice. Journal of Management Studies, 43(3), 641–53. Hara, N. and Schwen, T. M. (2006). Communities of practice in workplaces. Performance Improvement Quarterly, 19(2), 93–114. Hennebert, M. A. and Bourguignon, R. (2021). La gestion des ressources humaines à l’ère numérique: occasion stratégique ou risque de marginalisation? In J. Bernier (ed.), L’intelligence artificielle et les mondes du travail. Perspectives sociojuridiques et enjeux techniques, Presses Universitaires de Laval. Holubová, B. (2022). Algorithm management: awareness, risks and responses of the social partners. September 2022 report published by FES. Competence Centre Future of Work and UNI Europa. Huysman, M. and Elkjaer, B. (2006). Organizations as arenas of social worlds; towards an alternative perspective on organizational learning. In OLKC Conference at the University of Warwick, Coventry, March (pp. 20–6). Huzzard, T. (2000). Labouring to Learn: Union Renewal in Swedish Manufacturing. Boréa Bokförlag. Huzzard, T. (2001). Discourse for normalizing what? The learning organization and the workplace trade union response. Economic and Industrial Democracy, 22(3), 407–31. Huzzard, T. and Östergren, K. (2002). When norms collide: learning under organizational hypocrisy. British Journal of Management, 13(S2), S47–59. Hyman, R. (1997). Trade unions and interest representation in the context of globalisation. Transfer: European Review of Labour and Research, 3(3), 515–33. Hyman, R. (2007). How can trade unions act strategically?. Transfer: European Review of Labour and Research, 13(2), 193–210. Illeris, K. (2009). A comprehensive understanding of human learning. In Contemporary theories of learning, Routledge, (pp. 1–14). Ingham, M. (2015). L’apprentissage organisationnel dans les coopérations. Revue française de gestion, 41(253), 55–81. Jarrahi, M. H., Newlands, G., Lee, M. K., Wolf, C. T., Kinder, E., and Sutherland, W. (2021). Algorithmic management in a work context. Big Data & Society, 8(2), 20539517211020332. Jolly, C. and Naboulet, A. (2017). Mutations digitales et dialogue social. Note de Synthèse France Stratégie. Kellogg, K. C., Valentine, M. A., and Christin, A. (2020). Algorithms at work: the new contested terrain of control. Academy of Management Annals, 14(1), 366–410. Kelly, J. (1998). Rethinking Industrial Relations: Mobilisation, Collectivism and Long Waves. Routledge. Koenig, G. (2015). L’apprentissage organisationnel-Repérage des lieux. Revue française de gestion, 41(253), 83–95. Lansbury, R. D. and Bamber, G. J. (2013). Technological change, industrial relations and human resource management. In  G. J. Bamber and R. D. Lansbury (eds), New Technology, Routledge, pp. 3–37. Lave, J. and Wenger, E. (1991). Situated Learning: Legitimate Peripheral Participation. Cambridge University Press. Leonard, D. and Sensiper, S. (1998). The role of tacit knowledge in group innovation. California Management Review, 40(3), 112–32. Leroy, Y. (2020). Le comité social et économique face à l’intelligence artificielle. In M. Le Friant, P. Adam, and Y. Tarasewicz, Intelligence artificielle, gestion algorithmique du personnel et droit du travail-Les travaux de l’AFDT, Dalloz.

Algorithmic management from a ‘fault line’ to a frontline opportunity  91

Lévesque, C. and Murray, G. (2010). Understanding union power: resources and capabilities for renewing union capacity. Transfer: European Review of Labour and Research, 16(3), 333–50. Lévesque, C., Fairbrother, P., and Roby, N. (2020). Régulation du travail et de l’emploi à l’ère numérique: introduction. Relations industrielles/Industrial Relations, 75(4), 633–46. Lucio, M. M. (2006). Introduction: employment relations in a changing society. In E. Alonson and M. M. Lucio (eds), Employment Relations in a Changing Society: Assessing the Post-Fordist Paradigm, Palgrave Macmillan, pp. 3–17. March, J. G. and Simon, H. A. (1958). Organizations. John Wiley & Sons. Meijerink, J. and Bondarouk, T. (2021). The duality of algorithmic management: toward a research agenda on HRM algorithms, autonomy and value creation. Human Resource Management Review, 100876. Müller-Jentsch, W. (2018). Seven decades of industrial relations in Germany: stability and change through joint learning processes. Employee Relations, 40(40), 634–53. Murray, G., Lévesque, C., Morgan, G. D., and Roby, N. (2020). Disruption and re-regulation in work and employment: from organisational to institutional experimentation. Transfer: European Review of Labour and Research, 26(2), 135–56. Mustchin, S. (2012). Unions, learning, migrant workers and union revitalization in Britain. Work, employment and society, 26(6), 951–67. Nicolini, D. and Meznar, M. B. (1995). The social construction of organizational learning: conceptual and practical issues in the field. Human Relations, 48(7), 727–46. Nicolini, D., Gherardi, S., and Yanow, D. (eds) (2003). Knowing in Organizations: A Practice-based Approach. ME Sharpe. Northcott, J., Rogers, P., Knetsch, W., and De Lestapis, B. (1985). Microelectronics in Industry: An International Comparison: Britain, Germany, France (Vol. 635). Policy Studies Institute. Ponce Del Castillo, A. (2020). Labour in the age of AI: why regulation is needed to protect workers. ETUI Research Paper – Foresight Brief. Ponce Del Castillo, A. (2021). The AI regulation: entering an AI regulatory winter?. ETUI Policy Brief, 2021-06. Seifert, A. (2018). Employee data protection in the transnational company. In R. Blanpain, V. De Stefano, and F. Hendrickx (eds), Game Changers in Labour Law: Shaping the Future of Work, Alphen aan den Rijn, pp. 177–92. Sena, J. A. and Shani, A. B. (1999). Intellectual capital and knowledge creation: towards an alternative framework. In J. Liebowitz (ed.), Knowledge Management Handbook, CRC Press LLC, pp. 6.1–6.29. Senge, P. M. (1990). The Art and Practice of the Learning Organization (Vol. 1). Doubleday. Sfard, A. (1998). On two metaphors for learning and the dangers of choosing just one. Educational Researcher, 27(2), 4–13. Shibutani, T. (1955). Reference groups as perspectives. American Journal of Sociology, 60(6), 562–9. Simms, M. and Holgate, J. (2010). Organising for what? Where is the debate on the politics of organising?. Work, Employment and Society, 24(1), 157–68. Slee, M. (2014). Learning to navigate enterprise bargaining: the process of innovation and adaptation in the union movement. Labour & Industry: A Journal of the Social and Economic Relations of Work, 24(3), 217–34. Smith, H. (2021). The ‘indie unions’ and the UK labour movement: towards a community of practice. Economic and Industrial Democracy, 43(3), 1369–90. Smuha, N. A. (2021). From a ‘race to AI’ to a ‘race to AI regulation’: regulatory competition for artificial intelligence. Law, Innovation and Technology, 13(1), 57–84. Sprogoe, J. and Elkjaer, B. (2010). Induction – organizational renewal and the maintenance of status quo. Society and Business Review, 5(2), 130–43.

92  Research handbook on human resource management and disruptive technologies

Strauss, A. (1978). A social world perspective. Studies in Symbolic Interaction, 1, 119–28. Strauss, A. L. (1993). Continual Permutations of Action. Aldine de Gruyter. Todolí-Signes, A. (2019). Algorithms, artificial intelligence and automated decisions concerning workers and the risks of discrimination: the necessary collective governance of data protection. Transfer: European Review of Labour and Research, 25(4), 465–81. Voss, K. and Sherman, R. (2000). Breaking the iron law of oligarchy: union revitalization in the American labor movement. American Journal of Sociology, 106(2), 303–49. Walker, S., Lucio, M. M., and Trevorrow, P. (2007). Tracing trade union innovation: a framework for evaluating trade union learning projects in a time of complexity. Transfer: European Review of Labour and Research, 13(2), 267–90. Wallis, E., Stuart, M., and Greenwood, I. (2005). ‘Learners of the workplace unite!’ an empirical examination of the UK trade union learning representative initiative. Work, Employment and Society, 19(2), 283–304. Wenger, E., McDermott, R. A., and Snyder, W. (2002). Cultivating Communities of Practice: A Guide to Managing Knowledge. Harvard Business Press. Zoll, R. (2003). ‘Gewerkschaften als lernende Organisationen: Was ist eigentlich das problem?’, Gewerkschaftliche Monatshefte, 5, 315–21.

PART III HRM FOR DISRUPTIVE AND DISRUPTED ORGANIZATIONS

7. HRM systems and online labour platforms: survival of the (mis-)fittest? Anne Keegan and Jeroen Meijerink

INTRODUCTION Fit is an important and longstanding theme in HRM scholarship. Seminal HRM models propose the importance of fit between HRM practices and business strategies to enable organizations to hire, appraise and reward employees and consequently to shape workforce behaviour in support of strategic objectives. Creating fit between HRM practices and the external (usually competitive) environment is often called vertical fit, strategic fit or external fit (Miles and Snow, 1984). Another type of fit often discussed is how HRM practices ‘fit with and support each other’ (Baird and Meshoulam, 1988: 122). This is commonly referred to as internal fit or horizontal fit (Wood, 1999). Institutional fit refers to the fit between HRM practices and expectations from stakeholders. Influence exerted by stakeholders over organizations developing HRM practices raises questions of the perceived legitimacy of HRM practices, which in turn impacts on organizational legitimacy and survival (Wood, 1999; Paauwe, 2004). The importance of fit was elevated by the study of HRM systems that became prominent by the mid-2000s. Offering a holistic perspective on issues relating to the fit, theorists responded to perceived gaps in ideas on fit, including gaps related to the multidimensionality of fit as well as the importance of context and industry dynamics to questions of fit over time (Boxall, 1998). Scholars addressed the complex relationships between HRM practices and organizational outcomes by highlighting the roles of multiple actors involved in the design, execution and use of such practices. These included HRM practitioners (as designers), line managers (as implementers) and employees (as perceivers). Guided by an enriched theoretical repertoire incorporating sensemaking (Weick, 1995) and signalling (Guzzo and Noonan, 1994) theories, Bowen and Ostroff (2004) proposed that HRM practices, when bundled as coherent, visible and distinctive systems, provide organizations with the capacity to convey strategic meaning to employees. Using bundles of coherent practices to signal intentions to employees, and reinforcement by line managers and coworkers within units, facilitates employees to make sense of and define the psychological meaning of their work situation (Guzzo and Noonan, 1994). This allows employees to appropriately interpret strategic information conveyed in HRM practices. HRM systems have been widely studied to improve understanding of how practices contribute to the emergence of higher-level collective effects that translate into organizational performance outcomes (Delmotte et al., 2012; Den Hartog et al., 2013). 94

HRM systems and online labour platforms  95

Notwithstanding the extensive contributions of this line of scholarship, HRM systems approaches are somewhat limited in their focus on organizations with clear boundaries and standard employment relationships. Studies of HRM systems that are not based on more or less stable employment relationships, which are not focused on organizations with clear boundaries, and where HRM practices are designed and implemented by diffuse actors from across or outside organizational boundaries, are far less common. For example, the presence of line managers with formal reporting and relatively stable relationships with employees is a common assumption in HRM systems theorizing (Purcell and Hutchinson, 2007; Delmotte et al., 2012; Den Hartog et al., 2013). In this chapter, we discuss HRM systems in the context of OLPs (Online Labour Platforms) including Uber, Deliveroo and Fivver. OLPs rely on information technologies to match supply and demand for freelance labour. OLPs, through their use of algorithmic management, disrupt HRM systems and challenge the importance of fit. This disruption arises from business models where OLPs and workers do not have traditional employment relationships within clear organizational boundaries. Rather than fit constituting a core goal of HRM systems, OLPs undermine fit and engineer mis-fit. Our key contribution is to revisit the question of fit in HRM systems in the context of gig work and OLPs in order to explore the rationale of engineering mis-fit, and examine if aiming for fit is still the HRM cornerstone it once was in contexts like OLPs.

ENGINEERING MIS-FIT One of the central contributions of the turn to HRM systems in the mid-2000s was to increase understanding of the ways that organizations convey meaning to employees through (bundles of) HRM practices. Following Bowen and Ostroff (2004), consistently implemented coherent sets of HRM practices that fit well with organizational strategy ensure that employees, collectively within a unit, perceive the meaning of such practices as intended by employers. By enacting them in similar ways and linking these to strategic logics of the business, line managers can use bundles of practices that are internally aligned to, in effect, boost the strategic signal towards employees (Den Hartog et al., 2013) about what the organization seeks from them. Bundles of practices are more visible than single HR practices and more effective at conveying strategic meaning to employees (Buller and McEvoy, 2012). In contrast, OLPs contradict strong systems logic. Rather than designing and implementing coherent and mutually reinforcing HRM practices, OLPs engineer mis-fit through contradictory HRM practices that are deployed to manage gig workers. Mis-fit appears not by mistake, but by design. The answer to this lies in the importance of institutional (mis-)fit to the disruptive nature of platform business models and their approach to (avoiding) employment relationships.

96  Research handbook on human resource management and disruptive technologies

Institutional Mis-fit by Design While all types of fit are considered in HRM scholarship, institutional fit (e.g. Boon et al., 2009) has generally been less central than internal and external fit (Lewis et al., 2019). It is, however, critical to understanding HRM practices for gig workers. Although its centrality has weakened over the years as contingent forms of work increase (Cappelli and Keller, 2013; Bondarouk and Brewster, 2016), the standard employment relationship still has high levels of legitimacy and remains a dominant form for regulating relationships between workers and employers in many countries (Crouch, 2019). It is also central to HRM theorizing to the extent that HRM scholarship often appears synonymous with the study of practices to manage standard employment relationships (Meijerink and Keegan, 2019; Cross and Swart, 2021; Minbaeva, 2021). OLP business models, however, do not fit with prevailing institutions for managing workers and are directly premised on disrupting the employment relationship as an institution (Aloisi and De Stefano, 2020). A primary aim for HRM practices deployed within OLPs is to uphold freelance or independent contractor models while at the same time using algorithmic (HR) management (Meijerink et al., 2021a) to control workers in order to attain network effects (Katz and Shapiro, 1985) and control labour costs (Aloisi and De Stefano, 2020). OLPs seek carve-outs from employment legislation and develop arms-length relationships with workers, which convey that workers are autonomous even while they are core to value creation in platform ecosystems (Keegan and Meijerink, 2023). HRM systems in OLPs bear out this claim to worker autonomy to some extent. As previous research shows (e.g. Veen et al., 2020; Meijerink et al., 2021b), workers for Deliveroo, a prominent OLP in the food delivery sector, are free to join and leave the ecosystem and have significant job autonomy. Similar findings have been identified in studies of crowdworkers (e.g. Wood et al., 2019). High levels of work autonomy are conventionally viewed as belonging to commitment-oriented HRM bundles. Other HRM practices, including training practices offered to Deliveroo workers (albeit outsourced to third-party agencies), are also associated with commitment-oriented HRM bundles. To reinforce that workers are not employees, a key practice found in most OLPs is that workers have no contact with human managers. OLPs deploy algorithmic management (Rosenblat and Stark, 2016; Aloisi and De Stefano, 2022), which leads to the muting of worker voice (Gegenhuber et al., 2020; Keegan and Meijerink, 2022) by cutting off essential forms of voice, including the opportunity to discuss issues with direct managers using, for example, an ‘open door policy’ (Wilkinson and Fay, 2011: 68). Gegenhuber et al. (2020: 1496) also offer evidence of limited voice opportunities for platform workers because they face voice ‘regimes [that] equip crowdworkers with a microphone where the platform addresses individual crowdworkers and can balance and adjust crowdworker voice, if not mute it’. Gig workers do not have such managerial relationships or support. Their lack of employment contracts and hiring by app on a gig-by-gig basis (Kuhn and Maleki, 2017) preclude this. HRM practices are designed to show that workers are freelancers

HRM systems and online labour platforms  97

and should not think of themselves as attached to the platform in terms of a traditional employment relationship (Duggan et al., 2020). Accordingly, OLPs use HRM practices that reinforce the freelance status of gig workers and autonomy from line managers, including low levels of voice. Although algorithmic autonomy-related HRM practices are widespread in OLP HR systems, suggesting perhaps a tendency towards commitment-oriented HRM, OLPs also rely on HRM practices designed to restrict worker autonomy. OLPs subject gig workers to control/compliance-based HRM practices including piece-based pay and appraisal against strict pre-set standards (e.g. working during peak hours). These HRM practices service a key strategic aim – to achieve network effects that occur when ‘the value of membership to one user is positively affected when another user joins and enlarges the network’ (Katz and Shapiro, 1994: 94). Platform firms achieve network effects by attracting workers to the platform and leveraging this to attract requesters of labour services (individuals and businesses), which leads, in turn, to more workers and, in turn, more clients (Meijerink and Keegan, 2019). To gain network effects, platforms grapple with issues of trust by requesters on the platform. This leads platforms to establish and uphold quality standards aimed at workers that are linked with control/compliance-oriented HRM practices, including those already mentioned and others. Some platforms, for example, select workers stringently based on single and sometimes double layers of selection involving algorithmic selection followed by human selection (e.g. Handy in house-cleaning services or Topcoder in programming) (Waldkirch et al., 2021). Platforms use algorithmic surveillance and control (Newlands, 2021) to discipline workers (Jarrahi and Sutherland, 2019). Practices include deactivating their accounts or suspending workers who do not conform to the standards set by the platform and embedded in algorithms (Wood et al., 2019). Frequent changes to the terms of payment or levels of quality required, coupled with opaque algorithmic decision-making, creates precarity for gig workers, who learn to anticipate algorithmic requirements and engage in forms of self-surveillance and self-disciplining to keep access to clients (Bucher et al., 2021). Internal Mis-fit by Design As is evident from the preceding section on the engineering of institutional mis-fit by OLPS who both want to control workers and also want to uphold freelance models of employment, internal fit between platforms’ HRM practices is low. Empirical research by Waldkirch et al. (2021: 2645) shows that platform firms ‘employ a “hybrid HRM system” that blends elements from a control-based and a high-performance system’. The hybrid nature of practices is at odds with most HRM systems theorizing, which emphasizes consistency, coherence and clarity in bundles of mutually reinforcing HRM practices, the meaning of which are clearly signalled (Den Hartog et al., 2013). HRM practices embody OLPs’ contradictory aims in order to obfuscate the status of workers and to avoid the appearance of standard employment relationships (Frenken et al., 2020). Platforms can be said to deliberately leverage an autonomy paradox whereby they achieve network effects by using algorithmic management to

98  Research handbook on human resource management and disruptive technologies

exert control over workers (Veen et al., 2020) and, at the same time, offer freelance contracts to workers, refuse to engage them as employees under the law and deny in public documentation and declarations that they are employers. OLP HRM systems contradict the commitment to internal fit that is central to HRM scholarship based on strong HRM systems theorizing (Keegan and Meijerink, 2023). Internal fit is not a priority for platforms – paradoxes, contradictions, obfuscation and ambiguity in HRM practices are a priority, strategically leveraged to allow platforms to have their cake and eat it too (Meijerink et al., 2021b). Despite this, platforms convey what they want from workers effectively through algorithmic management. Algorithms convey the platform’s strategic goals to workers in terms of how they should treat requesters and behave as platform workers. This type of HRM approach disrupts the standard employment relationship – and thus creates both internal mis-fit and institutional mis-fit – as freelancers are controlled by platforms in ways that are similar to the management of standard employees but without offering gig workers the protections and rights that employees ought to enjoy. Although HRM practices are ambiguous and contradictory by design, the platform’s intentions for workers are signalled by algorithmic means. Non-conformance with algorithmic instruction can cost workers their access to work and wages. Excluded from protections associated with employment relationships, gig workers have less power and little recourse to appeal algorithmically imposed conditions of work (Veen et al., 2020), which can appear nonnegotiable (Rosenblat, 2018). External Fit as Dynamic and Unstable External fit is traditionally viewed as the fit between HRM practices and the strategic goals of the organization. Also called strategic or vertical fit, it is regarded as essential to ensure that HRM practices are not wasteful, avoid under- or over-investment in employees and ensure employees are focused on behaviours which realize strategic goals (Lepak and Snell, 2002). External fit between platform HRM practices and platform business strategy is prone to dynamism and instability linked with the ambiguous aims of platforms to control workers to achieve network effects while avoiding the appearance that workers are (or should be classified as) employees (Keegan and Meijerink, 2023). The goal of avoiding worker designation as employees – while platforms algorithmically control workers – often places platforms on a collision course with regulatory agencies, unions, national labour courts, activists and other actors ancillary to platform ecosystems (Frenken et al., 2020; Aloisi and De Stefano, 2020; Tassinari and Maccarrone, 2020; Meijerink et al., 2021b). External fit is prone to instability as platforms react to legal changes, threats of scrutiny from regulatory bodies and action by unions and activists (Prassl, 2018; Veen et al., 2020; Keegan and Meijerink, 2023). Platforms may temporarily change tactics by increasing/decreasing control over workers or by changing how they hire, appraise, train or remunerate workers. Although the strategy is usually fairly stable – to maintain the autonomy paradox – the practices deployed may be altered at very short notice as platforms respond tac-

HRM systems and online labour platforms  99

tically to threats or opportunities presented by the institutional context and ancillary actors to the platform ecosystem.

A RESEARCH AGENDA BASED ON (MIS-)FIT IN OLP HRM SYSTEMS Approaches to HRM in OLPs clearly differ from more traditional internal organizational and employee-oriented HRM systems. The first assumption that is challenged by OLPs is that HRM practices are rooted in internal fit, and that they will comprise coherent bundles of mutually reinforcing practices guided by a particular philosophy, e.g. high performance, high commitment, or control/compliance based, etc. Research shows that gig workers in OLPs are simultaneously subject to commitment and compliance/control-based HRM systems (Goods et al., 2019; Meijerink et al., 2021b; Waldkirch et al., 2021; Keegan and Meijerink, 2023). Gig workers in crowd-based platforms (e.g. Upwork, Fiverr) are tightly controlled through algorithm-based appraisal systems (Wood et al., 2019; Sutherland et al., 2020) and subject to electronic surveillance to monitor compliance (e.g. screenshots). In addition to these kinds of compliance-based HRM practices, platform workers are simultaneously subject to what are traditionally viewed as commitment-based HRM practices, including particularly extensive job autonomy. Proposition 1: Internal alignment among HRM practices is deliberately undermined by OLPs; they pursue hybrid HRM practices to uphold freelance contracting relationships while controlling workers using algorithms to achieve network effects. The power of OLPs to vertically align HRM practices with their strategic business goals (i.e. to maintain a freelance worker model and deploy algorithmic management to grow the platform) is mitigated by the fact that HRM practices reflect institutional mis-fit, making OLPS vulnerable to actions taken by other actors in response to this institutional disruption and the complexity flowing from it (Meijerink et al., 2021b). These actors include regulatory bodies, labour unions and activists who operate adjacent to platform ecosystems (Frenken et al., 2020; Tassinari and Maccarrone, 2020). These actors may seek to disrupt OLPs in implementing certain HRM practices based on algorithmic control and surveillance by taking court cases to challenge the legitimacy of OLPs’ business models. The influence of push-pull dynamics between core and ancillary ecosystem actors can, therefore, undermine OLPs’ efforts to achieve external or strategic fit. Proposition 2: External fit between HRM practices for gig workers and OLP strategy is unstable due to interactions between core actors and ancillary actors producing dynamic outcomes.

100  Research handbook on human resource management and disruptive technologies

A key distinction between conventionally theorized HRM systems and HRM systems in OLPs is that, in the former, workers inside organizational boundaries are core, and therefore subject to HRM systems framed by guiding principles and strategic goals (Lepak and Snell, 1999), while those workers ‘outside’ the organization are non-core, and subject to limited or no HRM at all (Kinnie and Swart, 2020). Gig workers, although core to the strategy of platform firms (e.g. for creating network effects, to meet requester needs and, thus, the continuation of offering intermediation services), are not employees (Kuhn and Maleki, 2017). Although this context is still considered somewhat extreme and institutionally volatile (Vandaele, 2018), the growing prevalence of deploying workers who do not have standard employment relationships recommends urgent attention from HRM scholars. Contexts where nonstandard work is common are largely overlooked from an HRM perspective (McKeown and Pichault, 2021). The practical implications of this, including the exclusion of freelancers from HRM analysis in areas like talent and career management (Dundon and Rafferty, 2018) are limiting the relevance and contribution of HRM scholarship (Cross and Swart, 2021; Minbaeva, 2021). HRM’s roots in RBV (resource-based view) theorizing (e.g. Wright et al., 1994) play a role here. Jacobides et al. (2018: 2270) make a pertinent observation: ‘Frameworks such as the RBV mostly concern themselves with owned resources. How should this perspective change when the resources exist not at the level of the firm, but at the level of the ecosystem?’ This question is highly pertinent for HRM in OLPs. However, it is also highly relevant to other contexts where HRM processes involve non-core workers and nonstandard employment relationships (Helfen et al., 2018). The current theorization of the HRM system as pertaining mainly to the management of employees means that HRM scholarship does not cover the whole workforce (Kuhn et al., 2021) but only those directly employed by the organization (Koene and Pichault, 2021). Future HRM research is needed that explicitly considers what the HRM system in any single organization includes/excludes, the rationale for this and the consequences for fit and mis-fit between HRM practices, HRM systems and organizational strategies. How HRM systems from different organizations overlap and influence each other’s many forms of ‘fit’ is also worthy of research (Kuhn et al., 2021). Proposition 3: Studies of fit between HRM practices, HRM systems and organizational strategies should be premised on considering non-core workers and workers with no contracts as well as those with standard employment relationships.

CONCLUSIONS AND RECOMMENDATIONS FOR ACTION Current theorizing on HRM systems is premised on assumptions that HRM practices should be aligned and coherent both internally and externally in order to support organizational strategy. Furthermore, mainstream HRM scholarship assumes such coherence and fit is required to convey meaning to employees about what the organization expects from them. These assumptions fail in the context of OLPs and their

HRM systems and online labour platforms  101

HRM systems. Understanding HRM practices in this context requires considering strategies of deliberate mis-fit as constituting organizational (or platform) aims and the knock-on effects for hybridity in internal HRM systems and sets of practices deployed to manage platform workers. OLPs use algorithmic HR management to convey their intentions to workers, and to encourage workers to conform to their expectations. Algorithmically managed practices for rewarding, disciplining and appraising workers, rather than shared social sensemaking between stable line manager-employee dyads supported by others (e.g. coworkers, senior managers etc.), are the main conduits for OLPs to control workers in service to their strategic goals – network effects and satisfying requesters. The engineering of institutional mis-fit and disruption also places OLPs on a collision course with external actors, creating dynamism and instability in strategic or external fit, and particularly institutional fit. By focussing on mis-fit, we are therefore better able to understand how HRM systems emerge, the dynamism that comes with OLPs’ approaches to managing workforces outside conventional organizational boundaries and the contradictory ways HRM practices are deployed to achieve complex institutional goals.

REFERENCES Aloisi, A. and De Stefano, V. (2020). Regulation and the future of work: the employment relationship as an innovation facilitator. International Labour Review, 159(1), 47–69. Aloisi, A. and De Stefano, V. (2022). Your Boss Is an Algorithm: Artificial Intelligence, Platform Work and Labour. Bloomsbury Publishing. Baird, L. and Meshoulam, I. (1988). Managing two fits of strategic human resource management. Academy of Management Review, 13(1), 116–28. Bondarouk, T. and Brewster, C. (2016). Conceptualising the future of HRM and technology research. The International Journal of Human Resource Management, 27(21), 2652–71. Boon, C., Paauwe, J., Boselie, P. and Den Hartog, D. (2009). Institutional pressures and HRM: developing institutional fit. Personnel Review, 38(5), 492–508. Bowen, D.E. and Ostroff, C. (2004). Understanding HRM–firm performance linkages: the role of the “strength” of the HRM system. Academy of Management Review, 29(2), 203–21. Boxall, P. (1998). Achieving competitive advantage through human resource strategy: towards a theory of industry dynamics. Human Resource Management Review, 8(3), 265–88. Bucher, E.L., Schou, P.K., and Waldkirch, M. (2021). Pacifying the algorithm – anticipatory compliance in the face of algorithmic management in the gig economy. Organization, 28(1), 44–67. Buller, P.F. and McEvoy, G.M. (2012). Strategy, human resource management and performance: sharpening line of sight. Human Resource Management Review, 22(1), 43–56. Cappelli, P. and Keller, J.R. (2013). Classifying work in the new economy. Academy of Management Review, 38(4), 575–96. Cross, D. and Swart, J. (2021). The (ir) relevance of human resource management in independent work: challenging assumptions. Human Resource Management Journal, 32(1), 233–46. Crouch, C. (2019). Will the Gig Economy Prevail?. John Wiley & Sons. Delmotte, J., De Winne, S. and Sels, L. (2012). Toward an assessment of perceived HRM system strength: scale development and validation. The International Journal of Human Resource Management, 23(7), 1481–506.

102  Research handbook on human resource management and disruptive technologies

Den Hartog, D.N., Boon, C., Verburg, R.M. and Croon, M.A. (2013). HRM, communication, satisfaction, and perceived performance: a cross-level test. Journal of Management, 39(6), 1637–65. Duggan, J., Sherman, U., Carbery, R., and McDonnell, A. (2020). Algorithmic management and app-work in the gig economy: a research agenda for employment relations and HRM. Human Resource Management Journal, 30(1), 114–32. Dundon, T. and Rafferty, A. (2018). The (potential) demise of HRM?. Human Resource Management Journal, 28(3), 377–91. Frenken, K., Vaskelainen, T., Fünfschilling, L., and Piscicelli, L. (2020). An institutional logics perspective on the gig economy. In I. Maurer, J. Mair, and A. Oberg (eds), Theorizing the Sharing Economy: Variety and Trajectories of New Forms of Organizing, Emerald Publishing Limited, pp. 83–105. Gegenhuber, T., Ellmer, M., and Schüßler, E. (2020). Microphones, not megaphones: functional crowdworker voice regimes on digital work platforms. Human Relations, 74(9), 1473–503. Goods, C., Veen, A. and Barratt, T. (2019). “Is your gig any good?” Analysing job quality in the Australian platform-based food-delivery sector. Journal of Industrial Relations, 61(4), 502–27. Guzzo, R.A. and Noonan, K.A. (1994). Human resource practices as communications and the psychological contract. Human Resource Management, 33(3), 447–62. Helfen, M., Schüßler, E. and Sydow, J. (2018). How can employment relations in global value networks be managed towards social responsibility?. Human Relations, 71(12), 1640–65. Jacobides, M.G., Cennamo, C., and Gawer, A. (2018). Towards a theory of ecosystems. Strategic Management Journal, 39(8), 2255–76. Jarrahi, M.H. and Sutherland, W. (2019). Algorithmic management and algorithmic competencies: understanding and appropriating algorithms in gig work. In International Conference on Information, Springer, March, pp. 578–89. Katz, M.L. and Shapiro, C. (1985). Network externalities, competition, and compatibility. The American Economic Review, 75(3), 424–40. Katz, M.L. and Shapiro, C. (1994). Systems competition and network effects. Journal of Economic Perspectives, 8(2), 93–115. Keegan, A. and Meijerink, J. (2022). Voice in platform-enabled gig work. In A. Wilkinson, T. Dundon, P. Mowbray, and S. Brooks (eds), Missing Voice? Worker Voice and Social Dialogue in the Platform Economy. Human Resource Management, Edward Elgar, pp. 153–74. Keegan, A. and Meijerink, J. (2023). Dynamism and realignment in the HR architecture: Online labor platform ecosystems and the key role of contractors. Human Resource Management, 62(1), 15–29. Kinnie, N. and Swart, J. (2020). Cross-boundary working: implications for HRM theory, methods, and practice. Human Resource Management Journal, 30(1), 86–99. Koene, B.A. and Pichault, F. (2021). Embedded fixers, pragmatic experimenters, dedicated activists: evaluating third-party labour market actors’ initiatives for skilled project-based workers in the gig economy. British Journal of Industrial Relations, 59(2), 444–73. Kuhn, K.M. and Maleki, A. (2017). Micro-entrepreneurs, dependent contractors, and instaserfs: understanding online labor platform workforces. Academy of Management Perspectives, 31(3), 183–200. Kuhn, K.M., Meijerink, J., and Keegan, A. (2021). Human resource management and the gig economy: challenges and opportunities at the intersection between organizational HR decision makers and digital labor platforms. In M. Buckley, J. Halbesleben, and A. R. Wheeler (eds), Research in Personnel and Human Resources Management, Emerald Publishing Limited, pp. 1–46.

HRM systems and online labour platforms  103

Lepak, D.P. and Snell, S.A. (1999). The human resource architecture: toward a theory of human capital allocation and development. Academy of Management Review, 24(1), 31–48. Lepak, D.P. and Snell, S.A. (2002). Examining the human resource architecture: the relationships among human capital, employment, and human resource configurations. Journal of Management, 28(4), 517–43. Lewis, A.C., Cardy, R.L. and Huang, L.S. (2019). Institutional theory and HRM: a new look. Human Resource Management Review, 29(3), 316–35. McKeown, T. and Pichault, F. (2021). Independent professionals as talent: evidence from individual views of working as a contractor. Human Resource Management, 60(2), 313–28. Meijerink, J. and Keegan, A. (2019). Conceptualizing human resource management in the gig economy: toward a platform ecosystem perspective. Journal of Managerial Psychology, 34(4), 214–32. Meijerink, J., Boons, M., Keegan, A. and Marler, J. (2021a). Algorithmic human resource management: synthesizing developments and cross-disciplinary insights on digital HRM. International Journal of Human Resource Management, 32(12), 2545–62. Meijerink, J., Keegan, A., and Bondarouk, T. (2021b). Having their cake and eating it too? Online labor platforms and human resource management as a case of institutional complexity. The International Journal of Human Resource Management, 32(19), 4016–52. Miles, R.E. and Snow, C.C. (1984). Designing strategic human resources systems. Organizational Dynamics, 13(1), 36–52. Minbaeva, D. (2021). Disrupted HR?. Human Resource Management Review, 31(4), 100820. Newlands, G. (2021). Algorithmic surveillance in the gig economy: the organization of work through Lefebvrian conceived space. Organization Studies, 42(5), 719–37. Paauwe, J. (2004). HRM and Performance: Achieving Long-term Viability. Oxford University Press. Prassl, J. (2018). Humans as a Service: The Promise and Perils of Work in the Gig Economy. Oxford University Press. Purcell, J. and Hutchinson, S. (2007). Front-line managers as agents in the HRM-performance causal chain: theory, analysis and evidence. Human Resource Management Journal, 17(1), 3–20. Rosenblat, A. (2018). Uberland: How Algorithms are Rewriting the Rules of Work. University of California Press. Rosenblat, A. and Stark, L. (2016). Algorithmic labor and information asymmetries: a case study of Uber’s drivers. International Journal of Communication, 10, 3758–84. Sutherland, W., Jarrahi, M. H., Dunn, M. and Nelson, S. B. (2020). Work precarity and gig literacies in online freelancing. Work, Employment and Society, 34(3), 457–75. Tassinari, A. and Maccarrone, V. (2020). Riders on the storm: workplace solidarity among gig economy couriers in Italy and the UK. Work, Employment and Society, 34(1), 35–54. Vandaele, K. (2018). Will trade unions survive in the platform economy? Emerging patterns of platform workers’ collective voice and representation in Europe. Emerging Patterns of Platform Workers’ Collective Voice and Representation in Europe (June 19, 2018). ETUI Research Paper-Working Paper. Veen, A., Barratt, T., and Goods, C. (2020). Platform-capital’s ‘app-etite’ for control: a labour process analysis of food-delivery work in Australia. Work, Employment and Society, 34(3), 388–406. Waldkirch, M., Bucher, E., Schou, P.K., and Grünwald, E. (2021). Controlled by the algorithm, coached by the crowd – how HRM activities take shape on digital work platforms in the gig economy. The International Journal of Human Resource Management, 32(12), 2643–82. Weick, K.E. (1995). Sensemaking in Organizations (Vol. 3). Sage.

104  Research handbook on human resource management and disruptive technologies

Wilkinson, A. and Fay, C. (2011). New times for employee voice?. Human Resource Management, 50(1), 65–74. Wood, A.J., Graham, M., Lehdonvirta, V., and Hjorth, I. (2019). Good gig, bad gig: autonomy and algorithmic control in the global gig economy. Work, Employment and Society, 33(1), 56–75. Wood, S. (1999). Human resource management and performance. International Journal of Management Reviews, 1(4), 367–413. Wright, P.M., McMahan, G.C. and McWilliams, A. (1994). Human resources and sustained competitive advantage: a resource-based perspective. International Journal of Human Resource Management, 5(2), 301–26.

8. Five decades of leadership and ‘disruptive’ technology: from e-leadership and virtual team leadership to current conversations on digital leadership Robin Bauwens and Laura Cortellazzo

INTRODUCTION Digital technologies (DTs), which encompass social, mobile, analytics, cloud and internet of things, currently present a disruptive force across multiple organisational levels (Sebastian et al., 2017). At the organisational level, DTs disrupt customer demands, business infrastructures and organisational boundaries (Vial, 2019). At team level, DTs transform work processes and erode existing hierarchies (Liao, 2017). At individual employee level, they alter productivity, but also workloads, work-home boundaries and skill requirements (Bauwens et al., 2020; Bauwens et al., 2021b). As several organisations undergo digital transformations and navigate the complexity of DTs – with mixed results – questions are raised over which success factors contribute to such processes (Weber et al., 2022). In response, an increasing number of scholars argue that leadership, defined as a process of influence in the interest of ‘facilitating individual and collective efforts to accomplish shared objectives’ (Yukl, 2010, p. 8), is critical to successfully manage the challenges associated with DTs (Avolio et al., 2014; Cortellazzo et al., 2019; Porfírio et al., 2021). However, do traditional leadership approaches account for situations where work and work processes are influenced or mediated by DTs? Alternatively, do they need to be adjusted – or even replaced – by other approaches? These and other questions have given rise to a five-decade-long scholarly conversation along the lines of concepts like ‘e-leadership’, ‘virtual team leadership’ and recently ‘digital leadership’. While each of these concepts reflect a desire to unravel what constitutes effective leadership against the background of DTs challenges in organisations (Torre and Sarti, 2020), different foci, shifts in meaning and the highly fragmented, multidisciplinary nature of this conversation – with contributions from information systems, organisational psychology and management – has made this a complex literature to traverse (Avolio et al., 2014; Cortellazzo et al., 2019). As a consequence, technology is not well-integrated in leadership literature and the impact of these studies remains limited as practice surpasses theory (Van Wart et al., 2019). Accordingly, the aim of this chapter is to present to the reader a brief overview of the conversation on leadership and technology, of which a graphical depiction is presented in Figure 8.1. We start from what we believe are the intellectual roots of 105

106  Research handbook on human resource management and disruptive technologies

the conversation: studies on computer-mediated groups examining the boundary conditions of traditional leadership approaches. Subsequently, we discuss the emergence of e-leadership and its consolidation in the debate on virtual team leadership. The discussion proceeds alongside a greater awareness of context, which led to the concept of digital leadership, with a more macro-orientation entering the conversation. We end our overview by discussing the impact of Covid-19, which has created a novel emphasis on virtual teams due to working from home mandates, before presenting the implications for future research and practice.

Figure 8.1

Historical overview of the conversation on leadership and technology

THE EVOLUTION OF THE CONVERSATION The Intellectual Roots The intellectual roots of conversations over what type of leadership is required for DTs can be traced back to the late 1970s when Strickland et al. (1978) investigated the difference between leadership in teleconferencing and face-to-face groups. It took another decade before research on leadership in ‘computer-mediated groups’ really took hold (e.g. Hiltz et al., 1991; Kahai et al., 1997; Sosik et al., 1997), and ‘virtual teams’ became the more dominant terminology (e.g. Jarvenpaa et al., 1998; Kayworth and Leidner, 2002). This demonstrates that, from its onset, research on leadership and DTs has always been strongly tied to the team context, which continues to be the case in recent work on virtual teams (e.g. Liao, 2017; Larson and DeChurch, 2020). The focus in these early works was largely on the application of traditional leadership styles and behaviours within a DTs context. For example, the contribution of transactional leadership and transformational (Sosik et al., 1997)

Five decades of leadership and ‘disruptive’ technology  107

or participative leadership (Kahai et al., 1997) to electronic group processes and outcomes. Other studies examined to what extent DTs were able to substitute certain aspects of leadership, by focusing on technological features, like feedback (Hiltz et al., 1991). The overall goal of these studies was, in the words of Sosik et al. (1997, p. 99), ‘to examine whether the boundary conditions in which leadership theories were conceived directly apply to interactions in computer-mediated environments’, the latter often referring to ‘group support systems’ (GSS), the precursors of today’s Zoom, Skype or Teams. Nevertheless, the generalizability of these studies remained limited, due to their predominant focus on lab experiments and student samples (Kahai et al., 2007). The Genesis of E-leadership The conversation was broadened by the seminal paper of Avolio et al. (2000), which is often regarded by scholars as the starting point for research on leadership and DTs. Lowe and Gardner (2000, p. 502) called the work: ‘an important genesis in elevating e-leadership, a topic that will likely take decades to explore’. While prior studies sometimes struggled with finding the right theoretical angle for ideas around leadership and technology to materialise, Avolio and his colleagues drew upon adaptive structuration theory (AST; DeSanctis and Poole, 1994), which emphasises that information technologies do not exist in a social vacuum but are in recursive interaction with the organisational structure, of which leadership is a key component. Since information technologies create changes in information access, sharing and work organisation, leaders are faced with different requirements and need to ensure that the necessary social structures are in place to support such changes. This is an important assertion, because it moves beyond examining the boundary conditions of traditional leadership approaches (cf. Sosik et al., 1997) by suggesting that traditional approaches might need to be complemented or even replaced to deal with DTs-induced requirements. To embody these novel leader requirements, the authors advanced the term ‘e-leadership’,1 referring to leadership in the context of work mediated by information technologies (Avolio et al., 2000). Between those arguing for a radical rethinking of leadership and those in favour of contextualising traditional leadership theories, Avolio and colleagues take a middle ground. ‘Even though many aspects of leadership will also remain the same’, the authors suggest that the use of information technology is likely to transform the patterns of behaviour that lead to effectiveness (Avolio et al., 2000, p. 660). E-leadership in Virtual Teams During the beginning of the twenty-first century, e-leader scholarship remained strongly connected to virtual team research. Scholars recognised that many organisations turned to teams and team-based working and were fascinated by the unique challenges of teams located in different parts of the world collaborating through information technologies (Avolio et al., 2000). Avolio et al. (2000) devoted a whole

108  Research handbook on human resource management and disruptive technologies

section of their paper to e-leadership in virtual teams. Accordingly, follow-up works focused largely on e-leaders’ role in fostering trust, communication, vision (i.e. transformational leadership) and shared decision-making (i.e. participative leadership) within virtual teams. For example, Zaccaro and Bader (2003) stated that trust is important because virtual teamwork leads to a sense of anonymity and loss of accountability that can hinder collaborative processes. Therefore, they argued, e-leaders need to develop trust in virtual teams by fostering information exchanges, interactions between team members and creating purpose and identity. Other examples include Zimmermann et al. (2008), Cordery et al. (2009) and Kayworth and Leidner (2002), who pointed to the importance of e-leader communication against the background of digital technology reducing social cues. To that end, these authors observed that e-leaders that were effective communicators and listeners were better able to align team members and address issues like misunderstandings or interpersonal conflicts. Some scholars continued along the path of the studies on computer-mediated groups by focusing on the boundary conditions of traditional leadership approaches, like transformational and shared leadership. Building on theories like media richness theory (Daft and Lengel, 1986), which draws on attention to how different information technologies affect the communication and tasks they are used for, these studies found that the effectiveness of some traditional leader approaches is enhanced within virtual teams. For example, Purvanova and Bono (2009) observed that transformational leadership was more influential in virtual than face-to-face teams. This dovetailed with studies that argue that DTs strengthen the need for a clear vision (e.g. Kelloway et al., 2003; Hambley et al., 2007) and that those leaders can appropriate DTs to express that vision (Avolio and Kahai, 2003). Carte et al. (2006) found that high-performing virtual teams were more likely to display shared or participative leadership in the form of monitoring and shared producing behaviours. Other illustrations include Yoo and Alavi (2004) and Kelly et al. (2008), showing that team members that engage in longer, task-related communication and that possess more technical proficiency are more likely to emerge as team leaders. In line with earlier works on participative leadership (e.g. Kahai et al., 1997), such studies demonstrate that DTs break down barriers related to communication and social hierarchy in favour of more informal forms of leadership, embedded collaborative processes between leaders and followers, as well as among followers themselves. Today, leadership in virtual teams is still a thriving field that contributes to the conversations on e-leadership and the broader field of leadership and DTs. This is demonstrated by recent overview works, which, besides providing a thorough state-of-the-art, sought a higher level of theoretical integration and advancement. For example, Liao (2017) has drawn attention to the multilevel structure of leadership and its outcomes in virtual teams (i.e. team level and individual level). Larson and DeChurch (2020) have addressed different relations between various leadership characteristics, styles or behaviours and DTs to demonstrate their recursive interactions. Furthermore, Purvanova and Kenda (2018, p. 780) have advanced a paradoxical virtual leadership model, arguing that e-leadership requires ‘perceiving the inter-

Five decades of leadership and ‘disruptive’ technology  109

relatedness between seeming contradictions, thinking holistically, and, ultimately, finding balance within virtuality’s numerous paradoxical tensions’. On an empirical level, topics like emergent and shared (Hoch and Dulebohn, 2017; Purvanova et al., 2021) or transformational leadership in virtual teams (Lauring and Jonasson, 2018) also remain popular to this day. The Contextual Turn and Further Refinement The second decade of the twenty-first century saw conversations on leadership and DTs seeking further refinement. In a novel paper, Avolio and his colleagues (2014) reflected upon their prior work (cf. Avolio et al., 2000) and the progress made to date by noting that DTs and their ‘appropriation at all levels of organisations and societies have far outpaced the practice and science of leadership’ (Avolio et al., 2014, p. 106). Accordingly, the authors argued that definitions of e-leadership would benefit from a stronger emphasis on context. That is, e-leadership not only refers to leadership mediated by DTs, but also to leadership embedded in a DTs context (i.e. the proximal team context and distal organisational context). Raising this issue of context helped to steer the conversation on leadership and technology, which had thus far largely committed itself to a specific part of the proximal leader-follower context (i.e. virtual teams) at the detriment of e-leader issues at the more distal organisational context. The coupling between e-leadership and virtual teams is also less prevent in Avolio et al. (2014), with ‘virtual groups’ regarded as only one micro-level locus of e-leadership. Works that subsequently committed themselves more to the distal organisational context include Cowan (2014) and Sharpp et al. (2019) focusing on e-leadership in nursing, Vermeulen et al. (2015) addressing e-leadership in education, and Li et al. (2016) examining how e-leadership can foster alignment between DTs and business strategy in small and medium-sized enterprises (SMEs). The works of Van Wart et al. (2017; 2019) and Roman et al. (2019), set against the context of the broader public sector, can also be counted to this group. These latter papers made an important contribution to the debate by proposing a competence-based model of e-leadership, the six e-competence (SEC) model. This represents a departure from previous studies, which have focused mostly on traditional leadership styles, and a first attempt to define those transformed patterns of behaviours that characterise e-leadership. The model suggests that is fundamental for e-leaders to be able to communicate through DTs (e-communication), create constructive virtual work environments (e-social) and teams (e-team) where people trust each other (e-trust), as well as to possess sufficient proficiency in change management (e-change) and DTs themselves (e-tech). Nevertheless, the emphasis is strongly on leaders’ communication skills (Van Wart et al., 2017; 2019). Follow-up work by Roman et al. (2019) developed an instrument to measure the SEC model, which currently constitutes one of the few validated scales to measure e-leadership.

110  Research handbook on human resource management and disruptive technologies

Digital Leadership Joining the Conversation Although the term is almost as old as e-leadership itself (cf. Fisk, 2002; Wilson, 2004), digital leadership has increasingly entered conversations on leadership and DTs, both in theory and practice, only in the last few years. Scholars like Torre and Sarti (2020) state that it is an attendant concept often used synonymously with e-leadership or virtual leadership. This seems to be true, both in research (e.g. in Cortellazzo et al., 2019; Roman et al., 2019; Torre and Sarti, 2020), as well as on a policy level (European Commission, 2015). However, existing definitions of digital leadership tend to adopt a more macro-level orientation concerned with DTs on a strategic level. For example, Wilson (2004) defines digital leadership as ‘leadership in the core sectors of the knowledge society’, El Sawy et al. (2016, p. 141) as ‘doing the right things for the strategic success of digitalisation for the enterprise and its business ecosystem’, and Larjovuori et al. (2016, p. 1444) as ‘leaders’ ability to create a clear and meaningful vision for the digitalisation process and the capability to execute strategies to actualize it’. In other words, terms like ‘e-leadership’ and ‘virtual team leadership’ have come to refer to what Avolio et al. (2014) calls the proximal context of e-leadership concerned with leaders’ behaviours, characteristics and skills committed to creating positive synergies between people and DTs. Conversely, digital leadership seems to correspond more with Avolio et al.’s (2014) view of the distal context of leadership concerned with aligning DTs with business imperatives. This latter interpretation is clearly illustrated in recent works like Porfírio et al. (2021) and McCarthy et al. (2021) that have sought to identify leader characteristics associated with a successful digital transformation at the organisation level. Despite these observations, the distinction between proximal and distal context is not always clear-cut. This is illustrated in the works of Zeike et al. (2019) and Petry (2018), authors that have taken a more micro-level approach to digital leadership. Zeike et al. (2019) built upon the definition of digital leadership by Larjovuori et al. (2016) to create a two-dimensional digital leadership scale, consisting of digital literacy and digital vision. Petry (2018) argued that DTs create a volatile, uncertain, complex and ambiguous (VUCA) environment for organisations and their leaders, necessitating a paradigm shift in leader requirements. Accordingly, he proposed a five-fold framework, the NOPA+ framework (i.e. network, openness, agility, participation, trust), which encompasses the five ‘novel’ requirements. Digital leaders need to seek and stimulate connections with other stakeholders (network), process and communicate information in a transparent, open-minded fashion (openness), stimulate the contributions of their followers (participation), be flexible, learning orientated and uncertainty resistant (agility) and foster confidence in themselves and their followers (trust). On the one hand, Petry’s (2018) model sounds promising. Dimensions like participation and trust dovetail with attention to these aspects in earlier studies (e.g. Kahai et al., 1997; Avolio et al., 2000; Zaccaro and Bader, 2003). Furthermore, Petry’s (2018) emphasis on agility and ambidexterity also aligns with studies that have adopted a leader complexity or paradox lens to look at leadership and DTs, like Purvanova and Kenda (2018) and more recently Weber et al. (2019; 2022).

Five decades of leadership and ‘disruptive’ technology  111

On the other hand, the above elements cannot hide the fact that the NOPA+ model is a bit elusive on its theoretical and empirical foundation. The model also shows strong overlap with the SEC model. For example, trust (trust, e-trust), change-readiness (e-change, agility), team participation (participation, e-social, e-team) and communication (openness, e-communication) feature centrally in both models. Therefore, one could ask to what extent a theoretical synthesis of both models is necessary or desired in the future. Both models can certainly learn from each other. The Present: Pushing Conceptual Boundaries and New Challenges In one of the most recent reviews of the field, Cortellazzo et al. (2019) took stock of the current conversation on leadership and technology and set out an agenda for research to come. Endorsing earlier work (i.e. Avolio et al., 2014), the authors confirmed that the conversation has developed along two lines: a macro-level perspective committed to leaders’ appropriation of DTs and their role in organisations undergoing digital transformation (distal context), as well as a micro-level perspective committed to leaders’ behaviours, characteristics and skills in relation to DTs and virtual teams (proximal context). This information led them to conclude that leaders need ‘to develop a combination of digital and human skills, mainly related to the ability to communicate effectively in a digitalised context, create cohesion between geographically distant followers, foster initiative and change attitudes, and deal with complex and fast problem solving’ (Cortellazzo et al., 2019, p. 47). However, the authors also made note of some caveats in the field. Especially the weak theoretical foundation – despite some occasional references to AST – and the lack of a clear definition, but also the scarce attention to networking, leader diversity and leader development, as well as leaders’ role in managing technology-induced ethical concerns. One of the main contributions of Cortellazzo et al. (2019) is its integration of the literature on e-leadership, virtual leadership and digital leadership. The authors seemed hesitant on committing to a certain terminology and opted for the more general ‘leadership in a digitalised world’ instead. Not long after, Torre and Sarti (2020) also made the case to consider these three concepts as part of the same intellectual conversation, albeit with slightly different foci. As the scholarship on leadership and DTs moves into its fifth decade, we see Covid-19 influencing the conversation. The recent working from home directives in multiple countries in response to such events led to the creation of many virtual teams across the globe. Questions like how to lead when followers only meet online invited several new scholars to the debate and gave a new impetus to the already significant popularity of leadership in virtual teams as a research topic (e.g. Chamakiotis et al., 2021; Chaudhary et al., 2022). According to Bauwens et al. (2021a), central in Covid-19 studies of leadership is a revaluation of context by focusing on specific affected settings (e.g. virtual teams) or affected sectors (e.g. healthcare, education), as well as more attention to specific issues like digital wellbeing. To an extent, this is also reflected in Covid-19 studies on leadership and DTs. For example, Spagnoli et al. (2020) examined the relationship of authoritarian leadership with workaholism and

112  Research handbook on human resource management and disruptive technologies

technostress among university teachers during the pandemic, while Salas-Vallina et al. (2022) looked at shared leadership and team resilience during Covid-19 in public healthcare. Venz and Boettcher (2021) showed that Covid-related work intensification could hinder the display of transformational leadership. Although Covid-studies of leadership and DTs investigate a unique setting characterised by a crisis and a sudden workplace change and organisational technology adoption (Mitchell, 2021), the future and the extent to which work will remain ‘hybrid’ or online in the coming years will tell whether Covid-studies of leadership and DTs will prove to be momentary or a lasting legacy. Currently, with hybrid work driving the workplace strategy and organisations still striving to find the optimal balance of workdays at home vs. in the office (PWC, 2021), eventual new challenges for e-leaders may arise. For instance, they will face the need to maximise the value created from both on-site and remote activities and to blend behaviours that drive effectiveness in virtual and non-virtual environments.

CONCLUSION AND RECOMMENDATIONS FOR ACTION Through its brief chronological overview, this chapter has highlighted how the conversation on leadership and DTs has evolved; from lab experiments in computer-mediated groups examining the boundary conditions of traditional leadership approaches, to leadership mediated by DTs in virtual teams and leadership that is embedded in DTs, driving the digital transformation at an organisational level. A couple of observations can be drawn from this overview. First, it seems the field currently suffers from a jingle-jangle fallacy, referring to the respective tendencies to consider two different concepts as similar due to a similar name (jingle fallacy) and to consider two similar concepts as discrepant due to different labels (jangle fallacy). The jingle fallacy originates from terms like e-leadership, virtual leadership and digital leadership being adopted by authors to mean different things. For example, to some scholars, e-leadership is traditional leadership within a DTs context (e.g. Hambley et al., 2007; Purvanova and Bono, 2009), while to others e-leadership refers to very specific competencies required to deal with DTs in organisations (e.g. Van Wart et al., 2017; Roman et al., 2019). In addition, there is a plethora of definitions currently in use for digital leadership. The jangle fallacy refers to the fact that, despite conceptual ambiguity and differences in emphasis, terms like e-leadership, virtual leadership and digital leadership belong to the same intellectual discussion (Torre and Sarti, 2020). Indeed, at the core of their respective literatures is a desire to understand what leader characteristics – old or new – are associated with successful outcomes for DTs in organisations. Accordingly, many scholars use the terms interchangeably (e.g. Roman et al., 2019; Torre and Sarti, 2020), while others avoid committing to a specific concept and adopt a more generalised terminology like ‘leadership in a digital world’ (Cortellazzo et al., 2019) or ‘leadership and DTs’ as we have done in this piece.

Five decades of leadership and ‘disruptive’ technology  113

A second observation is that much of what we know about leadership and DTs originates from research on virtual teams. While this is hardly a novel insight considering the intellectual roots of the field, it is important to reiterate this point because it seems earlier calls by Avolio et al. (2014) and Cortellazzo et al. (2019) to address the scant attention to the distal organisational context to the detriment of the proximal team context have largely been ignored. Indeed, leadership is a multilevel phenomenon (Batistič et al., 2017; Liao, 2017). Leadership in relation to DTs can take on a different meaning depending on whether one examines a team leader, a c-suite leader or even an informal emergent leader. Accordingly, different levels might translate into different behaviours, styles or competence to grow in importance relative to others. For example, e-leadership at organisational level might involve more strategic technology adoption characteristics to align DTs with organisational strategy (Liu et al., 2016), while e-leadership at team level might require more relational and team-oriented characteristics (Golden and Veiga, 2008). While we do see more attention to the distal organisational context in recent work (e.g. Porfírio et al., 2021; Nowacka and Rzemieniak, 2022), the large amount of Covid-19 research on virtual teams risks nullifying this catch-up (e.g., Chamakiotis et al., 2021; Chaudhary et al., 2022). A final observation is that, despite a lack of a common conceptualisation or uneven attention to different organisational levels, there are some ‘recurring leader characteristics’ throughout the conversation. Examples include leaders’ communication, trust, adaptability, vision and technical skills (cf. Kayworth and Leidner, 2002; Zaccaro and Bader, 2003; Liu et al., 2016; Zeike et al., 2019). As highlighted before, there is also a close similarity between the SEC and NOPA+ model, the former a competency-based model addressing e-leadership (Roman et al., 2019) and the latter a style-based based model centred around digital leadership. Such observations give credence to the possibility that the scholarly conversation on leadership and DTs might reach a ‘semantic and conceptual convergence’, resulting in a common, testable model on leadership and DTs. Research Recommendations While our narrative overview has already highlighted some deficiencies in the literature, we wish to stipulate three additional points that future research could address. A first recommendation concerns the limited attention to ethics in research on leadership and DTs (Avolio et al., 2014; Cortellazzo et al., 2019). While modern DTs already present some ethical challenges, like workplace monitoring, safe storage of information or blurring of work-life boundaries, this is exacerbated by more advanced technologies like artificial intelligence, smart technology or wearables that introduce ethical issues like employee privacy or unintentional algorithmic bias (Meijerink et al., 2021). This requires appropriate leader responses, in the form of ‘leadership skills to discern moral dilemmas, prioritise values, assess risks, protect privacy, and make ethical decisions’ (Lee, 2009, p. 462). However, aside from Lee

114  Research handbook on human resource management and disruptive technologies

(2009), Lin et al. (2020) and some minor attention by Roman et al. (2019), attention to ethical leadership is largely absent from research on leadership and DTs. Second, much of our knowledge on leadership and DTs originates from studies conducted in a Western, Global-North context. While this is a general critique of the leadership and management literature (Dar et al., 2021), it is particularly relevant to the current conversation given the rapid digitalising economy in countries like India, Indonesia and Vietnam, but also Kenya, Rwanda and Argentina (Chakravorti et al., 2020). For example, Belitski and Liversage (2019) have examined South African small and medium-sized enterprises and found that e-leadership in this context shows more commercial and knowledge-exploitation elements. Accordingly, we invite researchers to unravel the characteristics of e-leaders in newly digitalising economies. A final direction, inspired by HRM research, is to look beyond isolated leader characteristics, towards their interactions and joint effects with HRM practices and systems like training, performance management systems and high-performance work practices (HPWS). While this chapter has demonstrated a decent body of research devoted to leadership and DTs and other contributions in this volume have made a similar case for research on HRM and DTs, leadership and HRM largely remain separately studied phenomena (Leroy et al., 2018). Accordingly, the risk presents itself that this disparity is reproduced in work addressing DTs. We believe there are many commonalities in – and research on leadership and DTs can benefit from – works addressing line manager implementation of HRM. Indeed, just like HRM (cf. Nishii and Wright, 2008), DTs are part of an intended organisational strategy implemented by line managers in different segments of the organisation (Basselier et al., 2001; Prahalad and Krishnan, 2002). Therefore, we call for a people management agenda on DTs examining different forms of ‘leader-HRM’ fit in relation to DTs outcomes. For example, inspired by ability-motivation-opportunity (AMO) theory (Appelbaum et al., 2000), future research could devolve into how leaders’ skills, motives and possibilities align with HRM and DTs outcomes. Other studies could follow signalling theory (Connelly et al., 2011) to investigate how leadership and HRM jointly influence technology-related outcomes. This mutual influence is more evident in technology adoption matters, with social and institutional support being crucial in technology adoption theories (Venkatesh et al., 2003), and awareness and evaluation of technology, which can be promoted by training, gaining prominence in the latest refinements of this theory (Van Wart et al., 2017). The efficacy of e-leaders’ behaviours related to monitoring, controlling and empowering followers might also be highly influenced by the consistency of such behaviours with performance management HRM policies. Lessons for Practice In conclusion, organisations must realise that DTs do not occur in a social vacuum, but that their utility depends on the people involved, i.e. employees and their leaders. This renders DTs a people management challenge. Through a short narrative over-

Five decades of leadership and ‘disruptive’ technology  115

view, this chapter has highlighted that leadership matters for DTs in organisations. Accordingly, organisations should: ● Recognise that DTs pose a challenge to the traditional leadership repertoire but does not fully eradicate what we know about good leadership. ● Understand that DTs not only influence leadership, but leadership in turn also influences DTs choices. ● Foster leader styles, competencies and behaviours that centre on (digital) communication, building trust, adapting to novel circumstances, formulating an inspiring vision and technical proficiency. ● Invest in leader development of such styles, competencies and behaviours at both organisational and team level.

NOTE 1.

The term ‘electronic leadership’ actually predates Avolio et al. (2000). One of the first to mention this term is Kerr (1986).

REFERENCES Appelbaum, E., Bailey, T., Berg, P. B., Kalleberg, A. L., and Bailey, T. A. (2000). Manufacturing Advantage: Why High-Performance Work Systems Pay Off. Cornell University Press. Avolio, B. J. and Kahai, S. S. (2003). Adding the ‘e’ to leadership: how it may impact your leadership. Organizational Dynamics, 31(4), 325–38. https://​doi​.org/​10​.1016/​S0090​ -2616(02)00133​-X. Avolio, B. J., Kahai, S., and Dodge, G. E. (2000). E-leadership: implications for theory, research, and practice. The Leadership Quarterly, 11(4), 615–68. https://​doi​.org/​10​.1016/​ S1048​-9843(00)00062​-X. Avolio, B. J., Sosik, J. J., Kahai, S. S., and Baker, B. (2014). E-leadership: re-examining transformations in leadership source and transmission. The Leadership Quarterly, 25(1), 105–31. https://​doi​.org/​10​.1016/​j​.leaqua​.2013​.11​.003. Bassellier, G., Reich, B. H., and Benbasat, I. (2001). Information technology competence of business managers: a definition and research model. Journal of Management Information Systems, 17(4), 159–82. https://​doi​.org/​10​.1080/​07421222​.2001​.11045660. Batistič, S., Černe, M., and Vogel, B. (2017). Just how multi-level is leadership research? A document co-citation analysis 1980–2013 on leadership constructs and outcomes. The Leadership Quarterly, 28(1), 86–103. https://​doi​.org/​10​.1016/​j​.leaqua​.2016​.10​.007. Bauwens, R., Batistič, S., Kilroy, S., and Nijs, S. (2021a). New kids on the block? A bibliometric analysis of emerging COVID-19 – trends in leadership research. Journal of Leadership & Organizational Studies, 1548051821997406. https://​doi​.org/​10​.1177/​1548051821997406. Bauwens, R., Denissen, M., Van Beurden, J., and Coun, M. (2021b). Can leaders prevent technology from backfiring? Empowering leadership as a double-edged sword for technostress in care. Frontiers in Psychology, 12. https://​doi​.org/​10​.3389/​fpsyg​.2021​.702648. Bauwens, R., Muylaert, J., Clarysse, E., Audenaert, M., and Decramer, A. (2020). Teachers’ acceptance and use of digital learning environments after hours: implications for work-life

116  Research handbook on human resource management and disruptive technologies

balance and the role of integration preference. Computers in Human Behavior, 112, 106479. https://​doi​.org/​10​.1016/​j​.chb​.2020​.106479. Belitski, M. and Liversage, B. (2019). E-Leadership in small and medium-sized enterprises in the developing world. Technology Innovation Management Review, 9(1), 64–74. https://​doi​ .org/​10​.22215/​timreview/​1212. Carte, T. A., Chidambaram, L., and Becker, A. (2006). Emergent leadership in self-managed virtual teams: A longitudinal study of concentrated and shared leadership behaviors. Group Decision and Negotiation, 15, 323–43. https://​doi​.org/​10​.1007/​s10726​-006​-9045​-7. Chakravorti, B., Bhalla, A., and Chaturvedi, R. S. (2020). Which economies showed the most digital progress in 2020. Harvard Business Review, 18. Retrieved from: https://​hbr​.org/​ 2020/​12/​which​-economies​-showed​-the​-most​-digital​-progress​-in​-2020. Chamakiotis, P., Panteli, N., and Davison, R. M. (2021). Reimagining e-leadership for reconfigured virtual teams due to Covid-19. International Journal of Information Management, 60, 102381. https://​doi​.org/​10​.1016/​j​.ijinfomgt​.2021​.102381. Chaudhary, P., Rohtagi, M., Singh, R. K., and Arora, S. (2022). Impact of leader’s e-competencies on employees’ wellbeing in global virtual teams during COVID-19: the moderating role of emotional intelligence. Employee Relations. https://​doi​.org/​10​.1108/​ER​ -06​-2021​-0236. Connelly, B. L., Certo, S. T., Ireland, R. D., and Reutzel, C. R. (2011). Signaling theory: a review and assessment. Journal of Management, 37(1), 39–67. https://​doi​.org/​10​.1177/​ 0149206310388419. Cordery, J., Soo, C., Kirkman, B., Rosen, B., and Mathieu, J. (2009). Leading parallel global virtual teams: lessons from Alcoa. Organizational Dynamics, 38(3), 204–16. http://​doi​.org/​ 10​.1016/​j​.orgdyn​.2009​.04​.00. Cortellazzo, L., Bruni, E., and Zampieri, R. (2019). The role of leadership in a digitalized world: a review. Frontiers in Psychology, 1938. https://​doi​.org/​10​.3389/​fpsyg​.2019​.01938. Cowan, L. D. (2014). E-leadership: leading in a virtual environment – guiding principles for nurse leaders. Nursing Economics, 32(6), 312. Daft, R. L. and Lengel, R. H. (1986). Organizational information requirements, media richness and structural design. Management Science, 32(5), 554–71. https://​doi​.org/​10​.1287/​mnsc​ .32​.5​.554. Dar, S., Liu, H., Martinez Dy, A., and Brewis, D. N. (2021). The business school is racist: act up!. Organization, 28(4), 695–706. https://​doi​.org/​10​.1177/​1350508420928521. DeSanctis, G. and Poole, M. S. (1994). Capturing the complexity in advanced technology use: adaptive structuration theory. Organization Science, 5(2), 121–47. https://​doi​.org/​10​.1287/​ orsc​.5​.2​.121. El Sawy, O. A., Kræmmergaard, P., Amsinck, H., and Vinther, A. L. (2016). How LEGO built the foundations and enterprise capabilities for digital leadership. MIS Quarterly Executive, 15(2). European Commission (2015). E-leadership: Digital Skills for SMEs. Retrieved from http://​ eskills​-scale​.eu/​fileadmin/​promote/​documents/​eleadership​_digital​_skills​_fin​.pdf. Fisk, P. (2002). The making of a digital leader. Business Strategy Review, 13(1), 43–50. Golden, T. D. and Veiga, J. F. (2008). The impact of superior–subordinate relationships on the commitment, job satisfaction, and performance of virtual workers. Leadership Quarterly, 19(1), 77–88. https://​doi​.org/​10​.1016/​j​.leaqua​.2007​.12​.009. Hambley, L. A., O’Neill, T. A., and Kline, T. J. (2007). Virtual team leadership: the effects of leadership style and communication medium on team interaction styles and outcomes. Organizational Behavior and Human Decision Processes, 103(1), 1–20. https://​doi​.org/​10​ .1016/​j​.obhdp​.2006​.09​.004. Hiltz, S. R., Johnson, K., and Turoff, M. (1991). Group decision support: the effects of designated human leaders and statistical feedback in computerized conferences. Journal of

Five decades of leadership and ‘disruptive’ technology  117

Management Information Systems, 8(2), 81–108. https://​doi​.org/​10​.1080/​07421222​.1991​ .11517922. Hoch, J. E. and Dulebohn, J. H. (2017). Team personality composition, emergent leadership and shared leadership in virtual teams: a theoretical framework. Human Resource Management Review, 27(4), 678–93. https://​doi​.org/​10​.1016/​j​.hrmr​.2016​.12​.012. Jarvenpaa, S. L., Knoll, K., and Leidner, D. E. (1998). Is anybody out there? Antecedents of trust in global virtual teams. Journal of Management Information Systems, 14(4), 29–64. https://​doi​.org/​10​.1080/​07421222​.1998​.11518185. Kahai, S., Fjermestad, J., Zhang, S., and Avolio, B. (2007). Leadership in virtual teams: past, present, and future. International Journal of E-Collaboration, 3(1), 1–7. Kahai, S. S., Sosik, J. J., and Avolio, B. J. (1997). Effects of leadership style and problem structure on work group process and outcomes in an electronic meeting system environment. Personnel Psychology, 50(1), 121–46. https://​doi​.org/​10​.1111/​j​.1744​-6570​.1997​ .tb00903​.x. Kayworth, T. R. and Leidner, D. E. (2002). Leadership effectiveness in global virtual teams. Journal of Management Information Systems, 18(3), 7–40. https://​doi​.org/​10​.1016/​j​.hrmr​ .2016​.12​.010. Kelloway, E. K., Barling, J., Kelley, E., Comtois, J., and Gatien, B. (2003). Remote transformational leadership. Leadership & Organization Development Journal, 24(3), 163–71. https://​doi​.org/​10​.1108/​014377303104469589. Kelly, E., Davis, B., Nelson, J., and Mendoza, J. (2008). Leader emergence in an internet environment. Computers in Human Behavior, 24(5), 2372–83. https://​doi​.org/​10​.1016/​j​ .chb​.2008​.02​.013. Kerr, E. B. (1986). Electronic leadership: a guide to moderating online conferences. IEEE Transactions on Professional Communication, (1), 12–18. Larjovuori, R. L., Bordi, L., Mäkiniemi, J. P., and Heikkilä-Tammi, K. (2016). The role of leadership and employee well-being in organisational digitalisation. In T. Russo-Spena and C. Mele (eds), Annual RESER Conference, pp. 1159–72. Larson, L. and DeChurch, L. A. (2020). Leading teams in the digital age: four perspectives on technology and what they mean for leading teams. Leadership Quarterly, 31(1), 101377. https://​doi​.org/​10​.1016/​j​.leaqua​.2019​.101377. Lauring, J. and Jonasson, C. (2018). Can leadership compensate for deficient inclusiveness in global virtual teams?. Human Resource Management Journal, 28(3), 392–409. https://​doi​ .org/​10​.1111/​1748​-8583​.12184. Lee, M. R. (2009). E-ethical leadership for virtual project teams. International Journal of Project Management, 27(5), 456–63. https://​doi​.org/​10​.1016/​j​.ijproman​.2008​.05​.012. Leroy, H., Segers, J., Van Dierendonck, D., and Den Hartog, D. (2018). Managing people in organizations: integrating the study of HRM and leadership. Human Resource Management Review, 28(3), 249–57. https://​doi​.org/​10​.1016/​j​.hrmr​.2018​.02​.002. Li, W., Liu, K., Belitski, M., Ghobadian, A., and O’Regan, N. (2016). e-Leadership through strategic alignment: an empirical study of small-and medium-sized enterprises in the digital age. Journal of Information Technology, 31(2), 185–206. https://​doi​.org/​10​.1057/​jit​.2016​ .10. Liao, C. (2017). Leadership in virtual teams: a multilevel perspective. Human Resource Management Review, 27(4), 648–59. https://​doi​.org/​10​.1016/​j​.hrmr​.2016​.12​.010. Lin, W. L., Yip, N., Ho, J. A., and Sambasivan, M. (2020). The adoption of technological innovations in a B2B context and its impact on firm performance: an ethical leadership perspective. Industrial Marketing Management, 89, 61–71. https://​doi​.org/​10​.1016/​j​.indmarman​ .2019​.12​.009. Lowe, K. B. and Gardner, W. L. (2000). Ten years of the leadership quarterly: contributions and challenges for the future. Leadership Quarterly, 11(4), 459–514. https://​doi​.org/​10​ .1016/​S1048​-9843(00)00059​-X.

118  Research handbook on human resource management and disruptive technologies

McCarthy, P., Sammon, D., and Alhassan, I. (2021). Digital transformation leadership characteristics: a literature analysis. Journal of Decision Systems, 1–31. https://​doi​.org/​10​.1080/​ 12460125​.2021​.1908934. Meijerink, J., Boons, M., Keegan, A., and Marler, J. (2021). Algorithmic human resource management: synthesizing developments and cross-disciplinary insights on digital HRM. International Journal of Human Resource Management, 32(12), 2545–62. https://​doi​.org/​ 10​.1080/​09585192​.2021​.1925326. Mitchell, A. (2021). Collaboration technology affordances from virtual collaboration in the time of COVID-19 and post-pandemic strategies. Information Technology & People, ahead-of-print. https://​doi​.org/​10​.1108/​ITP​-01​-2021​-0003. Nishii, L. H. and Wright, P. (2008). Variability at multiple levels of analysis: implications for strategic human resource management. In D. B. Smith (ed.), The People Make the Place, Erlbaum, pp. 225–48. Nowacka, A. and Rzemieniak, M. (2022). The impact of the VUCA environment on the digital competences of managers in the power industry. Energies, 15(1), 185. https://​doi​.org/​10​ .3390/​en15010185. Petry, T. (2018). Digital leadership. In K. North, R. Maier, and O. Haas (eds), Knowledge Management in Digital Change, Springer, pp. 209–18. Porfírio, J. A., Carrilho, T., Felício, J. A., and Jardim, J. (2021). Leadership characteristics and digital transformation. Journal of Business Research, 124, 610–19. https://​doi​.org/​10​.1016/​ j​.jbusres​.2020​.10​.058. Prahalad, C. K. and Krishnan, M. S. (2002). The dynamic synchronization of strategy and information technology. MIT Sloan Management Review, 43(4), 24–34. Purvanova, R. K. and Bono, J. E. (2009). Transformational leadership in context: face-to-face and virtual teams. Leadership Quarterly, 20(3), 343–57. https://​doi​.org/​10​.1016/​j​.leaqua​ .2009​.03​.004. Purvanova, R. K. and Kenda, R. (2018). Paradoxical virtual leadership: reconsidering virtuality through a paradox lens. Group & Organization Management, 43(5), 752–86. https://​doi​ .org/​10​.1177/​1059601118794102. Purvanova, R. K., Charlier, S. D., Reeves, C. J. and Greco, L. M. (2021). Who emerges into virtual team leadership roles? The role of achievement and ascription antecedents for leadership emergence across the virtuality spectrum. Journal of Business and Psychology, 36, 713–33. https://​doi​.org/​10​.1007/​s10869​-020​-09698​-0. PWC (2021). It’s time to reimagine where and how work will get done. Available at https://​ www​.pwc​.com/​us/​en/​library/​covid​-19/​us​-remote​-work​-survey​.html. Roman, A. V., Van Wart, M., Wang, X., Liu, C., Kim, S., and McCarthy, A. (2019). Defining e-leadership as competence in ICT-mediated communications: an exploratory assessment. Public Administration Review, 79(6), 853–66. https://​doi​.org/​10​.1111/​puar​.12980. Salas-Vallina, A., Rofcanin, Y., and Las Heras, M. (2022). Building resilience and performance in turbulent times: the influence of shared leadership and passion at work across levels. BRQ Business Research Quarterly, 25(1), 8–27. https://​doi​.org/​10​.1177/​ 23409444211035138. Sebastian, I. M., Ross, J. W., Beath, C., Mocker, M., Moloney, K. G., and Fonstad, N. O. (2017). How big old companies navigate digital transformation. MIS Quarterly Executive, 16(3), 197–213. Sharpp, T. J., Lovelace, K., Cowan, L. D., and Baker, D. (2019). Perspectives of nurse managers on information communication technology and e-Leadership. Journal of Nursing Management, 27(7), 1554–62. https://​doi​.org/​10​.1111/​jonm​.12845. Sosik, J. J., Avolio, B. J., and Kahai, S. S. (1997). Effects of leadership style and anonymity on group potency and effectiveness in a group decision support system environment. Journal of Applied Psychology, 82(1), 89–103. https://​doi​.org/​10​.1037/​0021​-9010​.82​.1​.89.

Five decades of leadership and ‘disruptive’ technology  119

Spagnoli, P., Molino, M., Molinaro, D., Giancaspro, M. L., Manuti, A., and Ghislieri, C. (2020). Workaholism and technostress during the COVID-19 emergency: the crucial role of the leaders on remote working. Frontiers in Psychology, 11, 3714. https://​doi​.org/​10​.3389/​ fpsyg​.2020​.620310. Strickland, L. H., Guild, P. D., Barefoot, J. C., and Paterson, S. A. (1978). Teleconferencing and leadership emergence. Human Relations, 31(7), 583–96. Torre, T. and Sarti, D. (2020). The ‘way’ toward e-leadership: some evidence from the field. Frontiers in Psychology, 2753. https://​doi​.org/​10​.3389/​fpsyg​.2020​.554253. Van Wart, M., Roman, A., Wang, X., and Liu, C. (2017). Integrating ICT adoption issues into (e-) leadership theory. Telematics and Informatics, 34(5), 527–37. https://​doi​.org/​10​.1016/​ j​.tele​.2016​.11​.003. Van Wart, M., Roman, A., Wang, X., and Liu, C. (2019). Operationalizing the definition of e-leadership: identifying the elements of e-leadership. International Review of Administrative Sciences, 85(1), 80–97. https://​doi​-​.org/​10​.1177/​0020852316681446. Venkatesh, V., Morris, M. G., Davis, G., and Davis, F. (2003). User acceptance of information technology: toward a unified view. MIS Quarterly, 27(3), 425–78. https://​doi​.org/​10​.2307/​ 30036540. Venz, L. and Boettcher, K. (2021). Leading in times of crisis: how perceived COVID-19-related work intensification links to daily e-mail demands and leader outcomes. Applied Psychology. https://​doi​.org/​10​.1111/​apps​.12357. Vermeulen, M., Van Acker, F., Kreijns, K., and Van Buuren, H. (2015). Does transformational leadership encourage teachers’ use of digital learning materials. Educational Management Administration & Leadership, 43(6), 1006–25. https://​doi​.org/​10​.1177/​1741143214535749. Vial, G. (2019). Understanding digital transformation: a review and a research agenda. The Journal of Strategic Information Systems, 28(2), 118–44. https://​doi​.org/​10​.1016/​j​.jsis​.2019​ .01​.003. Weber, E., Büttgen, M., and Bartsch, S. (2022). How to take employees on the digital transformation journey: an experimental study on complementary leadership behaviors in managing organizational change. Journal of Business Research, 143, 225–38. https://​doi​.org/​10​ .1016/​j​.jbusres​.2022​.01​.036. Weber, E., Krehl, E.-H., Buettgen, M., and Schweikert, K. (2019). The digital leadership framework: insights into new leadership roles facing digital transformation. Academy of Management Proceedings, 2019(1), Article e13650. https://​doi​.org/​10​.5465/​AMBPP​.2019​ .13650abstract. Wilson, E. J. (2004). Leadership in the digital age. In G. R. Goethals, G. Sorenson, and J. M. Burns (eds), Encyclopedia of Leadership, Sage, pp. 858–61. Yoo, Y. and Alavi, M. (2004). Emergent leadership in virtual teams: what do emergent leaders do?. Information and Organization, 14(1), 27–58. https://​doi​.org/​10​.1016/​j​.infoandorg​ .2003​.11​.001. Yukl, G. (2010). Leadership in Organizations. Pearson. Zaccaro, S. J. and Bader, P. (2003). E-leadership and the challenges of leading e-teams: minimizing the bad and maximizing the good. Organizational Dynamics, 31(4), 377–87. https://​ doi​.org/​10​.1016/​S0090​-2616(02)00129​-8. Zeike, S., Bradbury, K., Lindert, L., and Pfaff, H. (2019). Digital leadership skills and associations with psychological well-being. International Journal of Environmental Research and Public Health, 16(14), 2628. https://​doi​.org/​10​.3390/​ijerph16142628. Zimmermann, P., Wit, A., and Gill, R. (2008). The relative importance of leadership behaviours in virtual and face-to-face communication settings. Leadership, 4(3), 321–37. https://​ doi​.org/​10​.1177/​1742715008092388.

9. Human resource management and customer value in the digital economy: advancing a value co-creation perspective Jeroen Meijerink

INTRODUCTION The creation of value is a central concept in strategic human resource management (HRM) research. Following the notion that firms have to create value first, before capturing it (Bowman and Ambrosini, 2000; Priem, 2007), HRM researchers studied the impact of HRM activities on customer value as a necessary condition to value capture (Lepak et al., 2007). Here, customer value is referred to as the utility of a product or service in terms of its quality versus costs to a customer (Zeithaml, 1988; Lepak et al., 2007). Building on the adage that frontline employees create value for customers, empirical investigations have shown that HRM activities impact customer outcomes through shaping employee attributes like their service climate perceptions (Hong et al., 2013) and service-oriented behaviors (Liao and Chuang, 2004). Moreover, strategic HRM scholars argued and found that HRM activities first have to be positively experienced by employees before impacting customer value (Liao et al., 2009; Piening et al., 2013; Meijerink et al., 2021a). As a result, these findings reinforced the consensus among HRM researchers that employees’ perceptions, attitudes and behaviors play an important role in translating HRM activities into customer value (Nishii et al., 2008; Hong et al., 2013). Although important, the focus on employees as key players marginalizes the role of customers in HRM-customer value relationships. In fact, scholars have repeatedly noted that the role of the customer in HRM research is largely absent, and if included, customers are solely seen as evaluators of service outcomes (Bowen and Pugh, 2009; Subramony and Pugh, 2015). This is remarkable as customers are active players who co-create value with employees, as shown in other streams of literature, such as marketing and strategy (Priem, 2007; Lusch and Vargo, 2011). In fact, in today’s business reality, digital technologies blur organizational boundaries by reinforcing customers’ participation in value creation processes by means of peer-to-peer platforms, online self-service or smart devices (Bowen, 2016; Breidbach and Maglio, 2016; Meijerink and Keegan, 2019). These are the technological developments that require HRM research to enter customers fully into the equation. Accordingly, this chapter provides a conceptual turn towards a value co-creation perspective on (i) how employee and customer attributes link HRM and customer value and (ii) how the explanatory power of these attributes depends on selected disruptive technologies. 120

Human resource management and customer value in the digital economy  121

SERVICE-ORIENTED HRM SYSTEMS AND CUSTOMER VALUE Customer value reflects the utility of a product or service in terms of its quality versus costs to a customer (Zeithaml, 1988; Lepak et al., 2007). Research has demonstrated that HRM activities like staffing, training, appraisal, compensation and job design relate positively to customer value, its subdimensions (e.g. service quality) and outcomes (e.g. customer satisfaction) (Rogg et al., 2001; Liao et al., 2009; Hong et al., 2013). HRM researchers stressed the importance of studying bundles of HRM practices as they are more likely to bring about desired outcomes when implemented in an integrated fashion (Wright and McMahan, 1992; Arthur, 1994). Here, the service-oriented HRM system is seen as most conducive to creating customer value (Liao and Chuang, 2004; Hong et al., 2013). Service-oriented HRM systems strongly impact customer value as they include HR practices that aim for improving service quality for customers, such as training for service-related skills, rewarding excellent service provision and empowering employees to meet idiosyncratic customer needs (Babakus et al., 2003; Liao et al., 2009). Indeed, research has shown that service-oriented HRM systems are more strongly related with customer outcomes in comparison to general HRM systems that intend to realize generic states such as employees’ general abilities, commitment or involvement (Hong et al., 2013). As discussed below, this thinking requires a revision due to the disruptive nature of the digital economy.

DIGITAL ECONOMY AND VALUE CO-CREATION WITH CUSTOMERS In the literature, the digital economy is referred to as the economic activities that are enabled by the internet and related information and communication technologies (Tapscott, 1996; Orlikowski and Iacono, 2000; Brynjolfsson and Kahin, 2002). The internet-enabled technologies play a disrupting role as they have led to an increased participation of customers in value creation processes (Schumann et al., 2012; Moeller et al., 2013; Bowen, 2016; Breidbach and Maglio, 2016). For instance, customers rely on online self-services that allow them to configure offerings themselves by putting together modular product/service components (Schumann et al., 2012). Furthermore, recent technological developments empower customers to co-create value with employees. This involves the use of wearables (such as smartphones, smartwatches or Google glasses) and smart machinery (e.g. operation robots and production equipment that is connected to the internet), which can be accessed, monitored and/or repaired remotely by a service provider (Schumann et al., 2012; Wuenderlich et al., 2015). The most significant role change for customers is brought about by online platforms such as Uber and Airbnb (Kuhn and Maleki, 2017; Meijerink and Keegan, 2019). These platforms (e.g. Uber) enable customers (e.g. the owner of a car) to share their assets with others and thus become a service provider

122  Research handbook on human resource management and disruptive technologies

(e.g. a taxi driver) (Ritzer and Jurgenson, 2010). This suggests that customer value is created in the first place by and among the customers themselves. In the literature, the involvement of customers in value creation processes has been referred to as ‘value-in-use’ creation. This holds that ‘value creation is only possible when a good or service is consumed’, meaning that value is created when customers utilize a service or product to meet their needs (Gummesson, 1998: 247). The utility of a product or service is, therefore, emergent from the actions of customers, which is referred to as value-in-use (Vargo and Lusch, 2004). To take this central role in value creation processes, I argue that customers need so-called high-level ‘role readiness’ in terms of abilities, motivation and opportunities to create value-in-use. While customers can grow their role readiness themselves, they can also call in the help of employees who can add to customer value by fostering customers’ abilities, motivation and opportunities to create value-in-use (Bowen, 2016; Priem, 2007). In the literature, this has been referred to as value co-creation; i.e. a process of resource integration, collaboration and service exchanges, which take place in networks of economic actors, including employees and customers (Vargo and Lusch, 2008, 2016; Lusch and Vargo, 2011). It is the notion of value co-creation that opens the road for a study how employee and customer attributes mediate HRM-customer value relationships in cases where digital technologies increase the level of participation of customers in value creation processes. Co-Creation-Oriented HRM System for Employee Co-Creation Behaviors To outline how HRM contributes to value co-creation, I introduce the co-creation-oriented HRM system, that is, as a bundle of HRM practices that aim to induce employee behaviors that are instrumental in providing customers with the abilities, motivation and opportunities to create value-in-use. As such, the co-creation-oriented HRM system is a so-called strategically-targeted HRM system that aims for a specific strategic objective (Jackson et al., 2014). In this respect, the co-creation-oriented HRM system is similar to the service-oriented HRM system, which I discussed before, as both aim to enhance customer value and related customer outcomes. At the same time, they differ in how this achieved. Namely, whereas the service-oriented HRM systems are geared towards enabling and motivating employees to provide high-quality services to customers (Liao et al., 2009; Jiang et al., 2015), the co-creation-oriented HRM system aims to ensure that employees engage in value co-creation with customers. Provided that customers need the abilities, motivation and opportunities to create value-in-use, a co-creation-oriented HRM system consists of HRM practices that encourage employees to instill these role readiness states in customers (see Figure 9.1). Specifically, a co-creation-oriented HRM system includes the following HRM practices: ● Staffing. This involves the hiring of T-shaped professionals with deep occupational expertise into a single field of service delivery (Dellande et al., 2004). Deep

Human resource management and customer value in the digital economy  123

Figure 9.1

Conceptual framework on HRM and value co-creation in the digital economy

occupational expertise among frontline employees is needed as this has a positive spillover effect on the abilities of customers (Dellande et al., 2004). Furthermore, deep-level competences likely aid employees to better understand the ends towards which customers put products and services to use – thereby facilitating value co-creation. ● Training. Organizations can further train employee to provide customers with the abilities and opportunities to create value-in-use. A training that is likely to be effective is the one that develops employees’ knowledge and skills on the value proposition canvas. The value proposition canvas is a method that helps employees to identify so-called customers’ ‘pains’ and ‘gains’ in getting certain jobs done. In receiving value canvas training, employees learn to identify customers’ jobs-to-be-done and how to align value propositions with these jobs-to-be-done, to ensure that customers are provided with the opportunity to create value-in-use (Osterwalder et al., 2014). ● Performance management. Employees’ behaviors can also be encouraged through performance management activities (Jiang et al., 2012), which involves measuring customer value-in-use levels (Blocker, 2011) and revenue generated from customers, as research has shown that customer value equates with customers’ willingness to pay (Bowman and Ambrosini, 2000; Lepak et al., 2007; Priem, 2007). Next to showing to employees the importance of providing high-quality services, this also encourages employees to grow customers’ abilities and motivate customers to share information as this ultimately adds to customers’ value perceptions (Priem, 2007; Meijerink et al., 2016).

124  Research handbook on human resource management and disruptive technologies

● Compensation and benefits. Companies can offer benefits that incentivize employees to increase customers perceptions of value and/or customer willingness-to-pay. Besides tying employees’ pay to customer value reports, this can be achieved by gainsharing programs. Employees’ compensation in gainsharing programs (such as the Scalon plan) is based on the income derived from customers (Bowen and Lawler III, 1995). This likely motivates employees to grow customers’ abilities or provide high-quality value propositions as it improves customers’ willingness to pay as an indicator of customer value (Bowman and Ambrosini, 2000; Lepak et al., 2007; Priem, 2007). ● Job design. To encourage employees to grow the abilities, motivation and opportunities of customers, companies can implement broad job-design practices such as autonomy and job enrichment as these provide employees the leeway to share knowledge with customers and customize services to the unique situation of a customer (Liao et al., 2009). Creating teams is equally important for value co-creation as team work stimulates learning among employees and fosters an organizational climate that stresses the importance of skill development that motivates employees to grow and develop themselves, but also customers (Iqbal et al., 2015). In my view, these HRM activities are key since they drive value co-creation by fostering three selected employee behaviors that impact customers’ role readiness to create value-in-use (see Figure 9.1). ● Customer development. This refers to employees’ efforts to grow and foster customers’ abilities. Marketing research shows that employees do indeed educate customers and improve customers’ abilities, for example, by providing information, instructions and even formal training on how to use a provided service/ product (Hibbert et al., 2012; Bell et al., 2017). Here, digital technologies can measure customer behavior and thus provide input for feedback on how customers can improve their consumption behavior (Rafaelli et al., 2017). For example, wearables such as smart glasses can supply an immediate and 360 degree view of customers’ consumption activities to service employees, which allows them to provide instant feedback on how the customer can consume a product or service more effectively or efficiently. This is also called pragmatic learning (Marinova et al., 2017), where feedback is derived from customer-provided data to enable customers to learn how to make better use of the services or products provided. ● Customer engagement. This to the efforts of employees for motivating customers to take part in co-producing services. The co-production of services by customers mainly involves them sharing data and information on e.g. their jobs-to-be-done, preferences and needs that products and services can be tailored to (Grönroos, 2011). In today’s digital economy, this active participation by customers does not come automatically, as customers may lack technology readiness, that is, the motivation to embrace and use cutting-edge technologies that help in sharing information (Parasuraman and Colby, 2015). In fact, research

Human resource management and customer value in the digital economy  125

shows that customers are not willing to share their information via online technologies if they feel that this empowers the service provider to harm their interests (Parasuraman and Colby, 2015; Breidbach and Maglio, 2016). Therefore, it is important that employees build trust and motivate customers to share information for co-producing high-quality value propositions. ● Resource orchestration. This involves employees’ efforts to provide customers with the opportunities to create value-in-use. In the service management literature, this opportunity is seen to come in the form of a value proposition, which employees create by integrating resources coming from within the firm and outside from customers (Grönroos, 2011; Vargo and Lusch, 2016). In the digital economy, the internet allows value propositions to be customized to a customer’s context as online self-services, wearables, or apps offer data on customers’ personal preferences and resource integration behaviors to which value propositions can be fitted (Rust and Huang, 2014; Marinova et al., 2017). This involves the collection of ‘big’ data on large groups of customers along with ‘small’ data, which is local, interpersonal and intuition-based knowledge of a customer’s unique situation that is needed to offset the shortcomings of big data (e.g. limited capture of the richness of a customer’s context). These data need to be turned into information and integrated into the current competence base of the service provider to contribute to the actual development of a value proposition that is tailored to the customer’s unique situation (Lam et al., 2017). The Contingent Role of Digitized Service Types Next to enabling employees to co-create value with customers, internet-based technologies also shape HRM and customer value relationships by acting as the context against which these relationships emerge and materialize. More specifically, I expect that HRM-customer value relationships (and the mediating role of employees’ co-creation behaviors and customers’ role readiness) are contingent on the type of internet-enabled service provided by organizations (see Figure 9.1). Following Moeller et al. (2013), I argue that almost any type of internet-enabled service can be fitted into one of three offerings to customers: solution services, networking services and configuration services. Each service type can be provided with the help of the internet and related technologies, yet requires customers to adopt different roles (Moeller et al., 2013), which explains differences in the explanatory power of selected mediating mechanisms in co-creation-oriented HRM and customer value relationships. Internet-enabled solution services The solution service is based on the customer’s problem, that is, a perceived difference between a current and a desired state. It involves the service company engaging in activities to solve customer problems. As noted by Moeller et al. (2013), these services are characterized by an information asymmetry between the provider and customer. That is, customers have more information about their problems, whereas

126  Research handbook on human resource management and disruptive technologies

providers have the specialized competences to solve them. Information technologies help to overcome these information/competence asymmetries through supporting collaborative problem-solving. Illustrative examples are wearables – such as smartphones and smartwatches – that are installed with e.g. eHealth apps that collect information which a doctor can use to solve the client’s problem (Lam et al., 2017; Marinova et al., 2017). Other examples include so-called smart machinery, which can be monitored and repaired remotely via the internet, or surgery robots, which allow doctors to operate on patients located on the other side of the globe (Schumann et al., 2012; Wuenderlich et al., 2015; Habraken and Bondarouk, 2017). As such, in the case of solution services, the internet allows customers to share information about their problem with the service provider and offer the service provider direct access to the solution space. Given these characteristics, I expect that the ‘customer motivation’ mediating mechanism best explains the HRM-customer value relationship in cases where internet-enabled solution services are provided. Namely, customers are almost always expected to co-produce solution service offerings, as they have to share information about their problem to reduce the information asymmetry that characterizes solution services (Moeller et al., 2013). At the same time, motivating customers to share information has proven to be challenging. Namely, research conducted in the solution domain shows that customers often withhold information because they lack trust in the service provider (Breidbach and Maglio, 2016). This should not come as a surprise, since the technologies that collect and share this information are heavily embedded in the lives of customers and thus may be misused (Wuenderlich et al., 2015). Some digital technologies even attach to or act upon the bodies of customers to solve their problems, for example, when a surgeon who is located on the other side of the globe is operating on a patient using a surgery robot (Schumann et al., 2012). Accordingly, it is key that potential customer concerns with regards to trust, privacy and wellbeing are taken care of, as otherwise they might lack the motivation to co-produce services and share information. Taken together, I expect that fostering customer motivation is an important activity for employees and, thus, plays an important role in explaining HRM-customer value in cases where internet-enabled solution services are provided. Proposition 1: In the case of internet-enabled solution service provision, the mediating effect of employees’ customer engagement behaviors and customers’ motivation best explains the HRM-customer value relationship. Internet-enabled configuration services Configuration services are characterized by bundles of relatively standard products/ services that customers configure into a customized output (Moeller et al., 2013). Exemplary internet-enabled configuration services are online self-services that allow customers to configure services themselves, such as a holiday trip by selecting from and combining hotels, flights, and/or rental cars that are suggested by a travel

Human resource management and customer value in the digital economy  127

agent. The same applies to e.g. online music streaming services (e.g. Spotify) where customers combine standardized input (i.e. songs) into a customized outcome (i.e. a playlist that fits their needs and circumstances – e.g. listening to up-tempo music to get going in the morning or soothing music to settle down at night) (Rust and Huang, 2014). Given the active role that customers take in configuration services, we expect that the ‘customer abilities’ mediating mechanism best explains the HRM-customer value relationship in cases where internet-enabled solution services are provided. For configuration services, growing and sustaining customer abilities is important. Research has shown that of all types of services, customers experience the highest level of role demands when provided with configuration services (Moeller et al., 2013). This can be explained by the fact that customers are almost entirely self-reliant when operating e.g. online self-services for configuring customized services that fit their individual needs (McKee et al., 2006). Under these conditions, improving customers’ abilities is important as it helps them to deal with their high-level role demands. In support of these claims, research conducted in online self-service settings showed that the influence of customer abilities on customer value is strong (McKee et al., 2006; Van Beuningen et al., 2009) as it increases the ease of use of these online self-services (Zhao et al., 2008; Meijerink et al., 2016). It should therefore be no surprise that organizations which offer configuration services have (online) helpdesks to support customers and provide them with the necessary abilities to configure services themselves. Given the importance of customer abilities in configuration services, I expect that fostering customer abilities is a key activity for employees and thus, plays an important role in explaining HRM-customer value in cases where internet-enables configuration services are provided. Proposition 2: In the case of internet-enabled configuration service provision, the mediating effect of employees’ customer development behaviors and customers’ abilities best explains the HRM-customer value relationship. Internet-enabled networking service The final service type – networking services – involves offerings that serve to link distributed customers who want to engage in transactions and/or share assets. These assets can be tangible (e.g. a car or an apartment) or intangible (such as ideas, information or data). In the case of networking services, the company is the middle man, or broker, whose main tasks include establishing and maintaining the network, linking customers and monitoring customer membership behaviors (Moeller et al., 2013; Meijerink et al., 2021b). Customers, on the other hand, are tasked with interacting with other customers through peer communication and offering/purchasing services (Moeller et al., 2013). An illustrative and emerging example of internet-enabled networking services are online platforms such as Uber, Deliveroo and Airbnb, whose underlying business model is to charge a fee for connecting dispersed economic actors through the internet (Kuhn and Maleki, 2017; Keegan and Meijerink, 2022).

128  Research handbook on human resource management and disruptive technologies

In the case of networking services, I expect that the ‘customer opportunities’ mediating mechanism best explains the HRM-customer value relationship. This follows from the fact that customer value is primarily created among customers themselves as they mutually engage in transactions without the direct involvement of the platform firm (Ritzer and Jurgenson, 2010; Meijerink and Keegan, 2019). As a result, the primary service offered by online platform firms and their employees is providing customers the opportunity to create value-in-use by establishing the linkages that allow customers to transact with one another (Moeller et al., 2013). That is, the raison d’être of companies that offer networking services is the provision of access to complementary resources (possessed by other customers) that a focal customer can integrate into their value-in-use creation activities. To ensure that opportunities for value-in-use creation are always present, platform firms such as Uber, Deliveroo and Airbnb employ a legion of data scientists who make sure that the supply of complementary resources coming from other customers matches a focal customer’s demand (Rosenblat and Stark, 2016; Kuhn et al., 2021). In so doing, platform firms can draw on a wide range of data (e.g. customer behaviors, preferences, context, etc.) for creating value propositions in terms of suggesting service-providing customers that possess complementary resources that best fit the needs of the service-receiving customer. To turn this data into value propositions requires employees’ resource integration behaviors that translate customer data into information, which is then integrated into the current competence base of the platform firm in terms of e.g. algorithms to match customers (Lam et al., 2017). Given that configuration services, by means of integrating and leveraging customer data, link customers that provide complementary resources to one another, I expect that customers’ opportunities to create value-in-use plays a key role in explaining HRM-customer value in cases where internet-enables networking services are provided. Proposition 3: In the case of internet-enabled networking service provision, the mediating effect of resource orchestration behaviors and customers’ opportunities to create value-in-use best explains the HRM-customer value relationship.

CONCLUSION AND RECOMMENDATIONS FOR ACTION Although the internet empowers customers to take an active part in value creation processes, the existing HRM literature merely treats customers as passive evaluators of customer value, rather than as active players who collaborate and co-create value with employees (Bowen and Pugh, 2009; Subramony and Pugh, 2015). Therefore, to better understand HRM-customer value relationships in contexts where digital technologies empower customers and employees to co-create value, this chapter has proposed a co-creation perspective for HRM research by developing a conceptual framework that links HRM with customer value through the mediating roles of both employee and customer attributes. The validity of this co-creation perspective for HRM, however, extends beyond the digital economy and has broader implications for HRM research, which I discuss here.

Human resource management and customer value in the digital economy  129

First, HRM researchers traditionally studied employee attributes as mediating mechanisms in HRM-customer value relationships, while overlooking customer attributes (Bowen and Pugh, 2009; Subramony and Pugh, 2015). A co-creation perspective provides the possibility to also study customers’ attributes for further opening the black box between HRM and customer value. Incorporating both employees and customers in the equation essentially extends the repertoire/portfolio of mediating mechanisms in the relationship between HRM and customer value. Ultimately, this gives a broader and more complete picture of the conceptual mechanisms through which HRM relates with customer outcomes and, thus, provides HRM researchers the possibility to more fully explain HRM-customer outcome relationships. Second, a co-creation perspective also makes it possible to see customers as being directly involved in the adoption and implementation of HRM practices. Such direct forms of customer involvement in HRM activities are likely to occur in the case of internet-enabled networking services. More specifically, in online platforms such as Uber, Deliveroo or Airbnb, performance management practices are exclusively delegated to and implemented by customers (Rosenblat and Stark, 2016; Meijerink et al., 2021b; Keegan and Meijerink, 2022). Here, service-receiving customers are expected to appraise the performance of service-providing customers (and vice versa) using 5 star rating scales or so-called ‘likes’ (e.g. thumbs up/down) (Pavlou and Gefen, 2004). Research, however, shows that customers may not know what consequences their appraisal activities might have for their service-providing peers (e.g. they do not know that a 4 out of 5 star rating is a failing grade) (Rosenblat and Stark, 2016). Also, I foresee the risk that customers might experience survey/evaluation fatigue and thus become ambivalent in their evaluations when making frequent use of many platforms (e.g. Uber, Airbnb, Deliveroo). Provided that customers have the means to co-create HRM practices and their outcomes, this warrants future studies to examine which customer abilities, motivations and opportunities play a role in customers’ involvement in implementing the HRM practices that are devolved to them. Finally, a co-creation perspective may help to explain how HRM is linked with employee service-oriented behaviors. Key in HRM research is the idea that HRM has to offer value to employees as it signals the importance the organization places on offering high-quality services (Bowen and Ostroff, 2004; Bowen and Pugh, 2009) and/or taking care of employee needs, which employees reciprocate by working towards organizational goals (Liao et al., 2009). In fact, research shows the importance of ensuring that HRM practices offer value to employees as this has favorable spillover effects on customer outcomes (Bowen and Pugh, 2009; Liao et al., 2009; Hong et al., 2013; Subramony and Pugh, 2015). Seen from a co-creation perspective, the value of HRM to employees is likely created in-use and emergent from how employees engage with provided HRM practices (Meijerink and Bos-Nehles, 2017). In support of this idea, research has shown that more than 90% of the variance in employees’ perceptions of HRM value resides at the employee-level (Liao et al., 2009; Meijerink et al., 2016), which suggests that the value of HRM is mainly affected by employee-level attributes. Indeed, research has shown that the value of HRM is affected by employees’ usage of HRM practices (Bondarouk et al., 2017)

130  Research handbook on human resource management and disruptive technologies

as well as their HRM-related competences that provide employees the abilities to create value-in-use out of provided HRM services (Meijerink and Bondarouk, 2018). Furthermore, the value of HRM practices to employees is likely to be co-created by a network of HRM actors – such as supervisors, HR shared service centers, and HR business partners – who provide HRM practices out of which employees can create value-in use (Bondarouk et al., 2018; Meijerink and Bondarouk, 2018). Taken together, I propose that future studies can benefit from adopting a co-creation perspective to examine whether employees’ abilities (e.g. their HRM competences), motivation (e.g. willingness to make use of provided HRM practices) and opportunities (e.g. support of HRM actors who provide HRM value propositions) affect employees’ perceptions of HRM value as an important antecedent to their customer-directed behaviors and, ultimately, related customer outcomes.

REFERENCES Arthur, J. B. 1994. Effects of human resource systems on manufacturing performance and turnover. Academy of Management Journal, 37(3): 670–87. Babakus, E., Yavas, U., Karatepe, O. M., and Avci, T. 2003. The effect of management commitment to service quality on employees’ affective and performance outcomes. Journal of the Academy of Marketing Science, 31(3): 272–86. Bell, S. J., Auh, S., and Eisingerich, A. B. 2017. Unraveling the customer education paradox: when, and how, should firms educate their customers? Journal of Service Research, 20(3): 306–21. Blocker, C. P. 2011. Modeling customer value perceptions in cross-cultural business markets. Journal of Business Research, 64(5): 533–40. Bondarouk, T., Bos-Nehles, A., Renkema, M., Meijerink, J., and de Leede, J. 2018. Organisational Roadmap Towards Teal Organisations: Emerald Group Publishing. Bondarouk, T., Harms, R., and Lepak, D. 2017. Does e-HRM lead to better HRM service? The International Journal of Human Resource Management, 28(9): 1332–62. Bowen, D. and Pugh, S. D. 2009. Linking human resource management and customer outcomes. In J. Storey, P. Wright, and D. Ulrich (eds), The Routledge Companion to Strategic Human Resource Management: 502–18. New York: Routledge. Bowen, D. E. 2016. The changing role of employees in service theory and practice: an interdisciplinary view. Human Resource Management Review, 26(1): 4–13. Bowen, D. E. and Lawler III, E. E. 1995. Empowering service employees. MIT Sloan Management Review, 36(4): 73–84. Bowen, D. E. and Ostroff, C. 2004. Understanding HRM–firm performance linkages: the role of the ‘strength’ of the HRM system. Academy of Management Review, 29(2): 203–21. Bowman, C. and Ambrosini, V. 2000. Value creation versus value capture: towards a coherent definition of value in strategy. British Journal of Management, 11(1): 1–15. Breidbach, C. F. and Maglio, P. P. 2016. Technology-enabled value co-creation: an empirical analysis of actors, resources, and practices. Industrial Marketing Management, 56(1): 73–85. Brynjolfsson, E. and Kahin, B. 2002. Understanding the Digital Economy: Data, Tools and Research. Cambridge, MA: MIT Press. Dellande, S., Gilly, M. C., and Graham, J. L. 2004. Gaining compliance and losing weight: the role of the service provider in health care services. Journal of Marketing, 68(3): 78–91.

Human resource management and customer value in the digital economy  131

Grönroos, C. 2011. Value co-creation in service logic: a critical analysis. Marketing Theory, 11(3): 279–301. Gummesson, E. 1998. Implementation requires a relationship marketing paradigm. Journal of the Academy of Marketing Science, 26(3): 242–9. Habraken, M. and Bondarouk, T. 2017. Smart industry research in the field of HRM: resetting job design as an example of upcoming challenges. In T. Bondarouk, H. Ruel, and E. Parry (eds), Electronic HRM in the Smart Era: 221–59: Emerald Publishing Limited. Hibbert, S., Winklhofer, H., and Temerak, M. S. 2012. Customers as resource integrators: toward a model of customer learning. Journal of Service Research, 15(3): 247–61. Hong, Y., Liao, H., Hu, J., and Jiang, K. 2013. Missing link in the service profit chain: a meta-analytic review of the antecedents, consequences, and moderators of service climate. Journal of Applied Psychology, 98(2): 237–67. Iqbal, S., Toulson, P., and Tweed, D. 2015. Employees as performers in knowledge intensive firms: role of knowledge sharing. International Journal of Manpower, 36(7): 1072–94. Jackson, S. E., Schuler, R. S., and Jiang, K. 2014. An aspirational framework for strategic human resource management. The Academy of Management Annals, 8(1): 1–56. Jiang, K., Chuang, C.-H., and Chiao, Y.-C. 2015. Developing collective customer knowledge and service climate: the interaction between service-oriented high-performance work systems and service leadership. Journal of Applied Psychology, 100(4): 1089–106. Jiang, K., Lepak, D. P., Hu, J., and Baer, J. C. 2012. How does human resource management influence organizational outcomes? A meta-analytic investigation of mediating mechanisms. Academy of Management Journal, 55(6): 1264–94. Keegan, A. and Meijerink, J. 2022. Dynamism and realignment in the HR architecture: online labor platform ecosystems and the key role of contractors. Human Resource Management. Kuhn, K. M. and Maleki, A. 2017. Micro-entrepreneurs, dependent contractors, and instaserfs: understanding online labor platform workforces. The Academy of Management Perspectives, 31(3): 183–200. Kuhn, K. M., Meijerink, J., and Keegan, A. 2021. Human resource management and the gig economy: challenges and opportunities at the intersection between organizational hr decision-makers and digital labor platforms. Research in Personnel and Human Resources Management. Lam, S. K., Sleep, S., Hennig-Thurau, T., Sridhar, S., and Saboo, A. R. 2017. Leveraging frontline employees’ small data and firm-level big data in frontline management: an absorptive capacity perspective. Journal of Service Research, 20(1): 12–28. Lepak, D. P., Smith, K. G., and Taylor, M. S. 2007. Value creation and value capture: a multilevel perspective. Academy of Management Review, 32(1): 180–94. Liao, H. and Chuang, A. 2004. A multilevel investigation of factors influencing employee service performance and customer outcomes. Academy of Management Journal, 47(1): 41–58. Liao, H., Toya, K., Lepak, D. P., and Hong, Y. 2009. Do they see eye to eye? Management and employee perspectives of high-performance work systems and influence processes on service quality. Journal of Applied Psychology, 94(2): 371. Lusch, R. F. and Vargo, S. L. 2011. Service-dominant logic: a necessary step. European Journal of Marketing, 45(7/8): 1298–309. Marinova, D., de Ruyter, K., Huang, M.-H., Meuter, M. L., and Challagalla, G. 2017. Getting smart: learning from technology-empowered frontline interactions. Journal of Service Research, 20(1): 29–42. McKee, D., Simmers, C. S., and Licata, J. 2006. Customer self-efficacy and response to service. Journal of Service Research, 8(3): 207–20. Meijerink, J., Beijer, S. E., and Bos-Nehles, A. C. 2021a. A meta-analysis of mediating mechanisms between employee reports of human resource management and employee

132  Research handbook on human resource management and disruptive technologies

performance: different pathways for descriptive and evaluative reports? The International Journal of Human Resource Management, 32(2): 394–442. Meijerink, J. and Bos-Nehles, A. 2017. Toward a marketing perspective on how ‘active employees’ create valuable human resource management outcomes: the role of HRM consumption and psychological ownership. In C. Olckers, L. van Zyl, and L. van der Vaart (eds), Theoretical Orientations and Practical Applications of Psychological Ownership: 159–77. Cham, Switzerland: Springer. Meijerink, J. G. and Bondarouk, T. 2018. Uncovering configurations of HRM service provider intellectual capital and worker human capital for creating high HRM service value using fsQCA. Journal of Business Research, 82(1): 31–45. Meijerink, J. G., Bondarouk, T., and Lepak, D. P. 2016. Employees as active consumers of HRM: linking employees’ HRM competences with their perceptions of HRM service value. Human Resource Management, 55(2): 219–40. Meijerink, J. G. and Keegan, A. 2019. Conceptualizing human resource management in the gig economy: toward a platform ecosystem perspective. Journal of Managerial Psychology 34(4): 214–32. Meijerink, J. G., Keegan, A., and Bondarouk, T. 2021b. Having their cake and eating it too? Online labor platforms and human resource mangement as a case of institutional complexity International Journal of Human Resource Management, 32(19): 4016–52. Moeller, S., Ciuchita, R., Mahr, D., Odekerken-Schröder, G., and Fassnacht, M. 2013. Uncovering collaborative value creation patterns and establishing corresponding customer roles. Journal of Service Research, 16(4): 471–87. Nishii, L. H., Lepak, D. P., and Schneider, B. 2008. Employee attributions of the “why” of HR practices: their effects on employee attitudes and behaviors, and customer satisfaction. Personnel Psychology, 61(3): 503–45. Orlikowski, W. J. and Iacono, C. S. 2000. The truth is not out there: an enacted view of the ‘digital economy’. In E. Brynjolfsson and B. Kahin (eds), Understanding the Digital Economy: Data, Tools, and Research: 352–80. Cambridge, MA: MIT Press. Osterwalder, A., Pigneur, Y., Bernarda, G., and Smith, A. 2014. Value Proposition Design: How to Create Products and Services Customers Want. Hoboken: John Wiley & Sons. Parasuraman, A. and Colby, C. L. 2015. An updated and streamlined technology readiness index: TRI 2.0. Journal of Service Research, 18(1): 59–74. Pavlou, P. A. and Gefen, D. 2004. Building effective online marketplaces with institution-based trust. Information Systems Research, 15(1): 37–59. Piening, E. P., Baluch, A. M., and Salge, T. O. 2013. The relationship between employees’ perceptions of human resource systems and organizational performance: examining mediating mechanisms and temporal dynamics. Journal of Applied Psychology, 98(6): 926–47. Priem, R. L. 2007. A consumer perspective on value creation. Academy of Management Review, 32(1): 219–35. Rafaelli, A., Altman, D., Gremler, D., Grewal, D., Iyer, B., Parasuraman, A., and de Ruyter, K. 2017. Invited commentaries on the future of frontline research. Journal of Service Research, 20(1): 91–9. Ritzer, G. and Jurgenson, N. 2010. Production, consumption, prosumption: the nature of capitalism in the age of the digital ‘prosumer’. Journal of Consumer Culture, 10(1): 13–36. Rogg, K. L., Schmidt, D. B., Shull, C., and Schmitt, N. 2001. Human resource practices, organizational climate, and customer satisfaction. Journal of Management, 27(4): 431–49. Rosenblat, A. and Stark, L. 2016. Algorithmic labor and information asymmetries: a case study of Uber’s drivers. International Journal of Communication, 10(27): 3758–84. Rust, R. T. and Huang, M.-H. 2014. The service revolution and the transformation of marketing science. Marketing Science, 33(2): 206–21.

Human resource management and customer value in the digital economy  133

Schumann, J. H., Wünderlich, N. V., and Wangenheim, F. 2012. Technology mediation in service delivery: a new typology and an agenda for managers and academics. Technovation, 32(2): 133–43. Subramony, M. and Pugh, S. D. 2015. Services management research: review, integration, and future directions. Journal of Management, 41(1): 349–73. Tapscott, D. 1996. The Digital Economy: Promise and Peril in the Age of Networked Intelligence. New York: McGraw-Hill. Van Beuningen, J., De Ruyter, K., Wetzels, M., and Streukens, S. 2009. Customer self-efficacy in technology-based self-service: assessing between- and within-person differences. Journal of Service Research, 11(4): 407–28. Vargo, S. L. and Lusch, R. F. 2004. Evolving to a new dominant logic for marketing. Journal of Marketing, 68(1): 1–17. Vargo, S. L. and Lusch, R. F. 2008. Service-dominant logic: continuing the evolution. Journal of the Academy of Marketing Science, 36(1): 1–10. Vargo, S. L. and Lusch, R. F. 2016. Institutions and axioms: an extension and update of service-dominant logic. Journal of the Academy of Marketing Science, 44(1): 5–23. Wright, P. M. and McMahan, G. C. 1992. Theoretical perspectives for strategic human resource management. Journal of Management, 18(2): 295–320. Wuenderlich, N. V., Heinonen, K., Ostrom, A. L., Patricio, L., Sousa, R., Voss, C., and Lemmink, J. G. 2015. ‘Futurizing’ smart service: implications for service researchers and managers. Journal of Services Marketing, 29(6/7): 442–7. Zeithaml, V. A. 1988. Consumer perceptions of price, quality, and value: a means-end model and synthesis of evidence. The Journal of Marketing, 52(3): 2–22. Zhao, X., Mattila, A. S., and Eva Tao, L.-S. 2008. The role of post-training self-efficacy in customers’ use of self service technologies. International Journal of Service Industry Management, 19(4): 492–505.

PART IV TECHNOLOGY-DRIVEN CHANGES IN HRM PRACTICE

10. Is artificial intelligence disrupting human resource management? A bibliometric analysis Stefano Za, Alessandra Lazazzara, Emanuela Shaba and Eusebio Scornavacca

INTRODUCTION The widespread adoption of artificial intelligence (AI) is bringing to light a number of important questions within the realm of human resources management (HRM): How is AI disrupting HRM? More specifically, which HRM practice has been most affected by the adoption of AI? Although AI is increasingly becoming a phenomenon of interest, there is still limited knowledge on the degree and the scope of its influence on HRM functions (Castellacci and Viñas-Bardolet, 2019). Therefore, the aim of this chapter is to answer such questions in order to understand how HRM research is tackling the increasing uptake of AI-enabled HRM applications (Canals and Heukamp, 2020) and the concurrent revolution in how HRM activities are delivered. HRM functions such as recruiting, performance appraisal, learning and development and talent acquisition, once completely carried out by humans, are now being restructured with the help of computer-generated assistants (Sivathanu and Pillai, 2018; Nankervis et al., 2019). For example, algorithm-based selection tools are increasingly finding their way into companies’ recruitment and selection processes, thereby saving costs and time and increasing efficiency of said processes. Moreover, with the assistance of AI-driven HRM solutions, practitioners are not only able to reduce human error and the administrative burden, but also fulfil other business purposes, such as enabling better decision-making (Davenport, 2019). In fact, AI, robotics, and machine learning can detect patterns in HRM functions, such as the need for more effective recruiting and selection and providing real-time guidance to professionals (Cooke et al., 2019; Nankervis et al., 2019). Furthermore, the prescriptive algorithms that are intensively automating HRM activities not only forecast what can be done in different scenarios but are also able to select and execute a course of action without (much) involvement by a human decision-maker. Given such potentialities, the adoption of AI into HRM has not come without its paradoxes (e.g. augmentation versus automation). Other issues relating to data privacy, fairness and ethical principles, as well as concerns related to employee autonomy, are pressing challenges for AI-enabled HRM. While AI is increasingly becoming a phenomenon of interest in many areas of HRM, there is a quite limited body of research that examines its specific adoption and 135

136  Research handbook on human resource management and disruptive technologies

evolution within the HRM field (Van Esch et al., 2019; Zehir et al., 2020). Given the increasing uptake of AI-enabled HRM applications, it is important for HRM scholars to have a robust understanding of the emerging trends, scenarios and challenges in the multidisciplinary literature of AI-enabled HRM. Given these considerations, our aim in this chapter is to identify the depth, range, significant contributors to, dynamics, trends, and domains of intersection between AI and HRM. In order to achieve this goal, we adopted a bibliometric approach, investigating 157 papers extracted from the Scopus database, thereby providing an overview and analysis of the current state of the art of research in the field use and application of HRM. In addition, current gaps in existing research have been identified, and also a future research agenda on the adoption of AI in HRM is provided. This chapter is structured as follows. The next section describes key trends emerging from the digitalization and the use of AI as applied to the study and practice of HRM. Following that, we present the research methodology and the literature search protocol. An overall coverage of the results of the bibliometric analysis is then presented. Finally, the last section presents discussion of results, conclusions, limitations and avenues for future research. The Emerging of AI in e-HRM Research AI disruption is rapidly revolutionizing the various HRM functions, and has thereby triggered an exciting new stream of research on the effects of AI applications on HRM practices (Bondarouk et al., 2017; Cooke et al., 2019). The development of digital transformation, a phenomenon describing the interconnection of physical and cybernetic environments through the use of technologies such as cyber-physical systems, big data, and cloud computing (Sony and Naik, 2019; Matt et al., 2020), allows for real-time collection of digital data that come in large amounts and from a variety of sources (Crawshaw et al., 2020). Such developments have further intensified the automation of HRM processes, a process encapsulated under the umbrella term ‘e-HRM’ (Bondarouk et al., 2017). Thereby, over the last decade, e-HRM has been enriched with such concepts as cloud computing, big data analytics, social media, chatbots, gamification, the internet of things, robots, artificial intelligence, etc. These technologies have been extensively adopted in the traditional HRM functions of recruitment, selection (e.g. Kehoe et al., 2005; Stone et al., 2015), learning and development, performance evaluation and compensation (e.g. Dulebohn and Marler, 2005). With the accelerating development and wide application of various AI-enabled technologies, HRM scholars have started extensively discussing how the emerging advanced systems, such as machine learning (ML) as a subset of AI, which can become adaptive and self-learning (for a review, see Vrontis et al., 2021), are used in decision making in many HRM related practices, creating novel premises for the execution of HRM activities. AI is typically defined as the use of digital technology to create systems capable of autonomously performing tasks commonly thought to require human intelligence

Is artificial intelligence disrupting human resource management?  137

(Office for AI, 2019). Recent advances in AI have occurred in the field of machine learning (ML), where digital systems autonomously improve their performance at undertaking a specific task or tasks over time as the system learns through experience (Office for AI, 2019). Put simply, AI is a technology that transforms digital inputs such as numbers, text, images and audio into outputs that can either be a decision or a solution to a problem (Crawshaw et al., 2020). Moreover, AI technologies enable machines to perform tasks like humans for a wide range of activities in areas such as cognition, sensing and performing (Akerkar, 2019). For each of these areas, AI integrates several databases of knowledge and uses AI tools to aid in business decisions and problem-solving. For example, for cognitive tasks, natural language processing technologies and knowledge representation approaches are critical. For sensing, AI relies on AI tools such as computer vision and imaging technologies as well as media processing tools. Similarly, for performing, essential AI tools employed for developing AI applications for completing human tasks include machine learning and knowledge-based systems (Crawshaw et al., 2020). In addition, there is already evidence of AI activity in chatbots and humanoids that can also offer and engage with humans not only at the cognitive level, but also at the empathetic level. With the help of AI, machines are hence also able to learn with experience and accomplish human-like tasks.

RESEARCH METHOD This study adopts a bibliometric approach to examine the literature intersecting AI and HRM, aiming to introduce a systematic, transparent and reproducible review process (Lamboglia et al., 2020). Bibliometric methods are a useful aid in literature reviews, as they guide the researcher to the most influential works and map the research field without subjective bias even before reading begins (Zupic and Cater, 2015). Such methods have two main uses: performance analysis and science mapping (Cobo et al., 2011). The former investigates the research and publication performance of individuals and institutions. The latter seeks to discover the structures and dynamics of the scientific field under investigation (Zupic and Cater, 2015). For studying the literature on AI and HRM in the context of various organizations, our bibliometric analysis has been applied with the assistance of tools such as citation, co-citation count (CoC), keyword frequency (keyword audit – KA), and co-authorship (Cisneros et al., 2018). The more co-citations two documents receive, the higher their co-citation strength, and the more likely they are semantically related. In the KA, by counting the appearance frequency of keywords, hotspots of disciplines can be analyzed (Huai and Chai, 2016). Thus, a combination of CoC (semantic relationship of documents), KA (hot disciplines) and co-authorship analysis (collaborations within a field) helps to determine the most prevalent areas in this field. The bibliometric analysis was carried out using R and the bibliometric package (Cuccurullo et al., 2016).

138  Research handbook on human resource management and disruptive technologies

For a wide-ranging systematic review of literature, three main steps were taken: (i) data collection, which consists of data loading and converting, (ii) analysis, including descriptive and network matrix creation (bibliographic coupling, co-citation, collaboration and co-occurrence analysis), and (iii) visualization, which provides mapping of the results. Data Collection and Analysis The first step involved a comprehensive search through a wide research query on Scopus, which is the largest abstract and citation database of peer-reviewed scientific literature in social studies, covering over 20,000 major journals (Bartol et al., 2014). The process related to the selection of the research query began with a literature review of the cornerstone papers related to AI, robotics and digital technologies (Caputo et al., 2018). As both concepts, AI and HRM, are studied alongside – but often interchangeably with – related phenomena including big data (García-Arroyo et al., 2019), Industry 4.0, or aggregated under the umbrella term ‘digital HRM’ (Strohmeier, 2020). After several iterations aiming to define a research query as broadly as possible to catch all possible papers, the resulting query was: (‘artificial intelligence’ OR ‘robot’ OR ‘chatbot’ OR ‘AI-based interviewing’ OR ‘asynchronous video interview’ OR ‘robot-mediated interview’ OR ‘robotics’ OR ‘social agent’ OR ‘virtual agent’) AND (‘selection decision’ OR ‘selection procedure’ OR ‘human resources selection’ OR ‘personnel selection’ OR ‘job interview’ OR ‘hiring’ OR ‘recruitment’ OR ‘recruitment decision’ OR ‘recruitment procedure’). We then performed a full search of the selected terms in titles and abstracts. This step generated a total of 834 articles in Scopus. The subsequent step was eliminating the duplicates and all out-of-topic papers in Scopus. For this purpose, we reduced the data to 194 articles. To ensure that all out-of-topic papers were identified, we performed a manual refinement of the dataset by reading the title, abstract and keywords of the 194 papers. In this selection, we followed the two-level methodology proposed by Keupp et al. (2012) to reduce subjectivity biases and encourage transparency and validity of the method. The authors carried out this analysis independently of each other, by agreeing in advance on the inclusion/exclusion criteria (Tranfield et al., 2003). In particular, we excluded the papers focused on HRM topics or on the technology exclusively, without any reference to the link between them. At the end of this process, the sample was composed of 157 articles. To proceed to the data analysis phase, the dataset had to be further refined. Since most of our analyses are based on the authors’ keywords, a list of reliable keywords for each paper in the dataset was needed. As a first step, we integrated the dataset with some authors’ keywords defined in the original paper but not listed in the exported dataset. The following step consisted of homogenizing keywords used in the articles (e.g. using only a plural or singular form). The keyword analysis was performed separately by two researchers and the results were then compared. In case of divergences or doubts, a discussion with the three members of the research team was used to reach an agreement (Za et al., 2018). Finally, for the papers without any keywords defined

Is artificial intelligence disrupting human resource management?  139

by the authors (neither in the dataset nor in the original papers), we integrated that missing information, adding a list of keywords identified on the basis of the abstract and the content of the specific paper, respecting at the same time the result of the keywords homogenization process previously performed. Once the dataset was cleaned and refined, we performed the data analysis. The data analysis involved two steps: (1) the analysis of descriptive performance indicators (descriptive analysis) and (2) the analysis of the conceptual structure of the dataset, exploring the main themes discussed in the corpus, through co-word analysis adopting social network analysis tools (bibliometric analysis).

DESCRIPTIVE BIBLIOMETRIC ANALYSIS The bibliometric analysis is presented in the following subsections: the trend of the number of publications, the most cited papers and the average number of citations of papers published in a specific year, most productive authors, journals’ publishing activity and publishing activity by country. Publication by Year The first papers investigating the relationship between HRM practices and AI appeared in the research literature around 20 years ago (see Figure 10.1). The publication trend until 2017 was quite constant and low, yielding 25 publications in total. An increasing trend is observed in the past 5 years, within which 2021 was the most productive, with 51 papers. Thereby, the period between 2017 and 2021 represents almost 87% of the total volume of publications in the dataset. Most Productive Authors Figure 10.2 shows the authors’ production over the time and reports authors with at least two publications in the dataset. The size of the dot expresses the number of published articles, while the intensity of the color refers to the total citations per year. In terms of the number of articles contributed, Nawaz N. has the highest, contributing seven articles during the period 2019–21, followed by König C.J. with four publications. In terms of authors that have a high impact in the literature, Gebhard P., Andrè E., Damian I., Jones H., and Baur T. have 94 citations each. Finally, the literary production over the period 2017–21 confirms the growing interest over these years in the research topic. Publishing Activity by Country The relationship between HR practices and AI seems to be a topic discussed worldwide, as indicated in the list of 27 countries according to the affiliation of the authors in the dataset. We are able to identify what is the output from different countries in

140  Research handbook on human resource management and disruptive technologies

Figure 10.1

Number of publications per year since 2003

Figure 10.2

Most productive authors

Is artificial intelligence disrupting human resource management?  141

the dataset. We conducted this analysis considering the ‘most productive country by volume’ (Figure 10.3) and the ‘average article citation per country’. Figure 10.3 distinguishes papers where the authors are from the same country (single country publication – SCP) and papers whose authors come from different countries (multiple country publication – MCP). The papers with authors with a multiple country configuration are associated with the country of the corresponding author. On the basis of this analysis, publications on AI and HR practices come mainly from the USA with a greater number of publications than in other countries (16 papers), followed by far fewer publications from China (nine papers), India (seven papers) and Germany (five papers). Germany and the United Kingdom also seem to be the most collaborative countries (two MCPs). The figures also reveal that the most productive countries are hence not the most collaborative and, overall, the collaboration on such topics among researchers from different countries is low. This could be related to the fact that this research stream is still in its infancy, and could also denote a lack of cross-cultural comparison on different AI applications in HRM with a resulting high local focus of such studies.

Figure 10.3

Most productive countries

Exploring the Main Topics Investigated in the Dataset To explore the most relevant topics discussed in our dataset, we analyzed the most often recurring keywords used by authors. This analysis is aimed at providing insights regarding the content and the main issues linking AI to HRM practices discussed in

142  Research handbook on human resource management and disruptive technologies

the realm of 157 papers of our dataset. The topic analysis is articulated in three parts: (1) the trend of the most used keywords; (2) authors keywords co-occurrences; (3) thematic map, used for identifying and classifying the main themes discussed in the corpus. Most Used Keywords This analysis started with the distribution of the most used keywords (32) defined by the authors over the years (Figure 10.4). AI is the most frequent keyword, with 65 occurrences, followed by recruitment, with 56 occurrences, and HR management, with 20. Job Interview (19), Selection (15), HR (13), and chatbot (11) were keywords with more than 10 and less than 20 occurrences. While machine learning (nine), hiring (seven), AI recruitment (five), deep learning (five) digital recruitment (five), fairness (five), HR analytics (five), and robotics (five) are comprised between 10 and five occurrences.

Figure 10.4

The main keywords used in the dataset over the years

The frequency of the 32 most common keywords is intensified between 2019 and 2022. We can observe that ‘AI’ has been used more frequently after 2017, while ‘recruitment’ presents a number of frequencies distributed throughout the years. Moreover, the graph shows that, so far, AI has been mainly explored as related to recruitment and selection practices. However, an emerging trend is a focus on the

Is artificial intelligence disrupting human resource management?  143

relationship between AI and the adoption of HR analytics, big data, and algorithms in HRM, as well as the growing interest in ethical issues related to the adoption of AI in recruitment and selection. Author Keywords Co-occurrences The analysis of the connections among the main topics discussed in our dataset can provide further insights about the AI-enabled applications in HRM. Starting from the 32 most recurring authors’ keywords, the co-occurrences network was created (Figure 10.5). In the graph, keywords are the nodes and there is a tie between two of them if mentioned together in the same publication (co-occurrence); the thickness reflects the number of contributions in which the pair appears. Finally, some of the most frequently recurring keywords are not mentioned, as the isolated nodes (without any connections with the nodes in the same graph) are deleted. On the basis of their connections, it is possible to recognize five clusters: (1) Cluster 1 includes the most cited keywords, namely ‘AI’ and ‘recruiting’, and is focused on the adoption of AI-technologies (e.g. chatbot, natural language processing, robotic) in order to automatize recruitment processes. (2) Cluster 2 focuses more generally on HR management and the link with HR technology. (3) Cluster 3 focuses on Selection and applicants’ reactions to the adoption of AI technologies in the interview process. (4) Cluster 4 focuses more specifically on job interview and AI and related fairness perception from the applicant perspective. (5) Cluster 5 focuses on talent management and the adoption of AI and social media in order to attract potential talent from outside the organization. The recruitment of potential employees was the most discussed topic related to the adoption of AI for HRM purposes, followed by the assessment and choice of the right candidates to fit into the organization. However, it is interesting to observe that no HRM or information systems-related theories have received attention in terms of keywords. Thematic Map In order to investigate further the topics discussed in the AI and HRM literature, we elaborate a thematic map, as suggested by Cobo et al. (2011). A thematic map shows clusters of keywords and their interconnections resulting in main research themes within a specific research field. Clusters’ creation is based on an algorithm that takes into consideration the keyword co-occurrence in the dataset. Once the clusters are identified, two parameters are calculated: ‘density’ refers to the strength of internal ties among all keywords in the same cluster (thus highlighting the development of the specific theme represented by that cluster) and ‘centrality’ refers to the strength

144  Research handbook on human resource management and disruptive technologies

Figure 10.5

Authors’ keywords co-occurrence

of the external connections among different clusters (thus depicting the relevance of the research theme for the specific field of study). The combination of high and low values for density and centrality parameters makes it possible to define a strategic diagram distributing the research themes into four quadrants (Figure 10.6): (1) Themes in the upper-right quadrant are the so-called ‘motor-themes’, which are high both in density and centrality and are well established for the specific research field under analysis (i.e. AI in HRM) as well as being related externally to concepts applicable to other fields (e.g. digital HR, e-HRM). (2) Themes in the upper-left quadrant are the so-called ‘niche-themes’; they are considered well developed and specialized but have only marginal importance for the field. (3) Themes in the lower-left quadrant are considered marginal and weakly established and can indicate emerging or declining themes. (4) Themes in the lower-right quadrant are relevant for specific fields but are still in the process of development. These themes are both basic and transversal. The thematic map built on the current dataset indicates that recruitment is the first and foremost HRM function discussing AI. At least five themes may be considered as motor-themes within the AI for HRM literature and mostly refer to the advancement

Is artificial intelligence disrupting human resource management?  145

Figure 10.6

Thematic map of the relationship between AI and HRM practices

that AI offers in recruitment processes. For example, a number of HRM recruitment processes now use robotics process automation, wherein bots are dealing with low-end, predictable and routine external requests from job applicants or indeed answering routine internal HRM policy queries for employees (Crawshaw et al., 2020). Similarly, algorithm-based tools have thus increasingly found their way into companies’ recruitment and selection processes, thereby saving cost and time and increasing efficiency. With the assistance of AI-driven HRM solutions, HRM practitioners are able not only to reduce the human error and the administrative burden, but also to provide better forecasts. In fact, prescriptive algorithms that automate recruitment activities not only forecast what can be done in different scenarios, but also select and execute a course of action without (much) involvement by a human decision maker. Given the above potentialities, and despite the increasing popularity and usage of algorithm-based selection tools, AI acceptance is still in its infancy (Baxter, 2018; O’Donovan, 2019). In fact, as algorithmic HRM is shifting decision-making responsibility from human to machine by automating decision-making processes (Duggan et al., 2020), a number of important topics, such as data privacy, ethics, fairness, or employee autonomy have also increased the perceived complexity of AI-enabled applications for HRM. Several computer-aided tools assist in matching aspirants to a job, which also help to reduce recruiters’ workload. Such methods include software that uses learning-based techniques and algorithms to resume and implement the matching

146  Research handbook on human resource management and disruptive technologies

process (Vrontis et al., 2021). Another remarkable facet of AI-based classifications is the possibility of collecting information about candidates’ personality characteristics, which are highly critical in filling job vacancies (Suen et al., 2019). For example, during the hiring process, measures such as information tests, cognitive tests, personality tests, reference checks, structured/unstructured interviews and work samples can be applied with sophistication and in time, hinting at the candidate that can perform best in the appropriate job position (Mahmoud et al., 2019). The upper-left quadrant shows high density themes such as training and performance evaluation that are currently external and of limited importance for the field (low centrality). In this regard, some experimental evaluation techniques, such as 360-degree performance appraisal approaches, can be used more efficiently and progressively adopting intelligent decision support systems (IDSS) (Góes and De Oliveira, 2020). Another trending theme within research on training in the workplace relates to how learning can be supported by AI, and how AI tools bring complementary strengths to achieve educational outcomes. However, despite emerging AI-enabled applications supporting learning and development and performance assessment practices (Sitzmann and Weinhardt, 2019), it seems that the analysis of AI in HRM literature has not yet considered other HRM practices in addition to the well-established focus on recruitment and selection. In the lower-left quadrant are the emerging or declining themes. In this research, the theme related to the adoption of AI-driven recruitment assistant (chatbots) constitutes an emerging topic together with the adoption of AI-based robots undertaking tasks that were performed by humans (task-automation). Finally, the lower-right quadrant shows the themes that are general topics transversal to the different research areas of the field. In this quadrant, the appearing themes are mainly related to the three most general keywords adopted in the AI for HRM literature stream, namely AI, recruitment, and HR management. However, there is a specific theme related to applicants’ reactions and asynchronous video interview, which seems to be a well-established basic theme transversal to other fields. Indeed, the role of an applicant’s personality (e.g. introvert vs. extravert) on the effect of AI adoption in automated job interview has not been fully explored and there is a growing interest on that in the psychological literature (Acikgoz et al., 2020; Nørskov et al., 2020).

CONCLUSIONS AND RECOMMENDATIONS FOR ACTION This chapter has performed an investigation of the current state of the art of research on the use and application of AI-enabled applications in HRM functions. In order to achieve this goal, we performed a bibliometric analysis on a sample of 157 articles articulated in three steps: first, we performed a descriptive bibliometric analysis; second, we identified the main topics of our dataset through a co-word analysis; and third, we developed a thematic map in order to identify and classify the main themes in the dataset.

Is artificial intelligence disrupting human resource management?  147

The results of the descriptive analysis highlight a growing academic interest in this research topic over time and an exponential increase in the number of publications in the past 5 years. The analysis of contributing authors indicates the existence of an emerging global community interested in investigating the relationship between HRM practices and AI. The analysis of the most common keywords provided by the authors has established the dominance of research focusing on the use of AI for recruitment and selection purposes. Also, it was possible to observe that ‘AI’ has been used more frequently in the past 5 years, while ‘recruitment’ was evenly distributed throughout the 20-year timeframe. The longitudinal keyword analysis also showed a growing interest in research focusing on big data analytics and algorithms delegation as well ethical issues related to HRM. Five clusters emerged from the analysis of the connections among the main topics discussed in our dataset (co-occurrences network). The first cluster presented papers focused on the adoption of AI-technologies in order to automatize recruitment processes, while the second cluster was composed of articles focused more generally on HR management and the link with HR technology. The third cluster had its focus on the selection process and applicants’ reactions to the adoption of AI technologies in the interview process. On the other hand, the fourth cluster presented literature with a narrower focus, specifically exploring the use of AI for job interviews and perceptions of fairness from the applicant’s perspective. Finally, the fifth cluster explored issues related to talent management and the adoption of AI through social media to recruit potential candidates. The analysis of the thematic map has shown the emergence of at least five motor-themes within the AI for HRM literature. These themes were mostly related to the advancement of AI tools for recruitment and selection processes. Two peripheral niche themes emerged focusing on aspects of performance evaluation and training, reinforcing the well-established focus of the existing body of literature on the use of AI for recruitment and selection purposes. Hence, significant attention was given to how AI applications can serve recruiters, and as Galanaki et al. (2019) point out, studies are keeping pace with how the recruitment process can be enhanced with AI technological support. In addition, our analysis shows that the application of AI-driven recruitment assistants (chatbots) is an emerging theme along, with other issues of anthropomorphic forms of task-automation that enable candidates to develop a more personalized connection with recruiters (Bondarouk and Brewster, 2016; Upadhyay and Khandelwal, 2018). Nonetheless, while AI and recruitment appeared to be topics transversal to the different research areas of the field, issues specifically related to applicants’ perceptions toward automated job interviews seem to be in vogue. However, O’Donovan (2019), in providing an overview of the historical developments of human resource management (HRM) functions, argues that even if the use of AI in the interview process is becoming a more common tool for organizations to hire new employees, recruiting the most competent and best employees in the market remains a challenging task.

148  Research handbook on human resource management and disruptive technologies

The bibliometric analysis is also indicative of existing research gaps in the context of AI and HRM-related functions, such as cross-cultural management, compensation, employment relations, etc. In fact, authors such as Wu et al. (2012) have proposed the application of AI-based models that enable companies and individuals to make intelligent decisions in culturally diversified organizational activities, and resolve the intercultural problems encountered in a variety of authentic cross-cultural circumstances. Furthermore, the lack of theories either from HRM or the information systems literature was clearly evident in the papers analyzed. The present study has both theoretical and practical implications. From a theoretical perspective, it contributes to the existing literature, providing useful insights as to the link between AI and HRM. In addition, this literature analysis should provide a solid starting point for researchers interested in pushing forward the research agenda on the use of AI in HMR. Specifically, through the results of the cluster analysis, we were able to pinpoint themes that are already well-developed as well as areas of research that could be further explored. From a practical standpoint, our research provides a useful guide to sources of literature for managers as well as an outline of emerging issues in the field, as each identified cluster discusses issues that have a significant impact on HRM functions. While this chapter makes a contribution, it also presents limitations. The study focused on papers published only in journals indexed on Scopus. In order to enrich our dataset, it could be interesting to take into consideration multiple databases. Another important limitation is that we bounded our investigation to authors’ keywords only. It would be interesting to develop the research performing not only co-citation analysis, but also cross-citation analysis to investigate the evolution of the community and how and if each author affects the others and for which topics. Finally, collaboration among authors could be analyzed, such as the relationships within the knowledge-creation process of the specific research community.

REFERENCES Akerkar, R. (2019). Artificial Intelligence for Business. Cham: Springer. Acikgoz, Y., Davison, K.H., Compagnone, M., and Laske, M. (2020). Justice perceptions of artificial intelligence in selection. International Journal of Selection and Assessment, 28(4), 399–416. https://​doi​.org/​10​.1111/​ijsa​.12306. Bartol, T., Budimir, G., Dekleva-Smrekar, D., Pusnik, M., and Juznic, P. (2014). Assessment of research fields in Scopus and Web of Science in the view of national research evaluation in Slovenia. Scientometrics, 98(2), 1491–504. Baxter, M. (2018). Information-age. https://​www​.information​-age​.com/​business​-analytics​ -intelligence​-123477004. Accessed 26 June 2020. Bondarouk, T. and Brewster, C. (2016). Conceptualizing the future of HRM and technology research. The International Journal of Human Resource Management, 27(21), 2652–71. https://​doi​.org/​10​.1080/​09585192​.2016​.1232296. Bondarouk, T.V., Parry, E., and Furtmueller, E. (2017). Electronic HRM: four decades of research on adoption and consequences. The International Journal of Human Resource Management, 28(1), 98–131.

Is artificial intelligence disrupting human resource management?  149

Canals, J. and Heukamp, F. (2020). The Future of Management in an AI World: Redefining Purpose and Strategy in the Fourth Industrial Revolution. IESE Business Collection. Caputo, A., Marzi, G., Pellegrini, M.M., and Rialti, R. (2018). Conflict management in family businesses: a bibliometric analysis and systematic literature review. International Journal of Conflict Management, 29(4), 519–42. Castellacci, F. and Viñas-Bardolet, C. (2019). Internet use and job satisfaction. Computers in Human Behavior, 90, 141–52. https://​doi​-org​.pros2​.lib​.unimi​.it/​10​.1016/​j​.chb​.2018​.09​.001. Cisneros, L., Ibanescu, M., Keen, C., Lobato-Calleros, O., and Niebla-Zatarain, J. (2018). Bibliometric study of family business succession between 1939 and 2017: mapping and analyzing authors’ networks. Scientometrics 117(2), 919–51. https://​ doi​ -org​ .pros2​ .lib​ .unimi​.it/​10​.1007/​s11192​-018​-2889​-1. Cobo, M.J., López-Herrera, A.G., Herrera-Viedma, E., and Herrera, F. (2011). An approach for detecting, quantifying, and visualizing the evolution of a research field: a practical application to the fuzzy sets theory field. Journal of Informetrics, 5(1), 146–66. https://​doi​ .org/​10​.1016/​j​.joi​.2010​.10​.002. Cooke, F.L., Liu, M., Liu, L.A., and Chen, C.C. (2019). Human resource management and industrial relations in multinational corporations in and from China: challenges and new insights. Human Resource Management, 58(5), 455–71. https://​doi​-org​.pros2​.lib​.unimi​.it/​ 10​.1002/​hrm​.21986. Crawshaw, J., Budhwar, P., and Davis, A. (2020). Human Resource Management (third edition). Sage Publishing. Cuccurullo, C., Aria, M., and Sarto, F. (2016). Foundations and trends in performance management. A twenty-five years bibliometric analysis in business and public administration domains. Scientometrics, 108(2), 595–611. Davenport, T. (2019). Can we solve AI’s ‘trust problem’? MIT Sloan Management Review. 60(3), 1–5. Duggan, J., Sherman, U., Carbery, R., and McDonnell, A. (2020). Algorithmic management and app-work in the gig economy: a research agenda for employment relations and HRM. Human Resource Management Journal, 30(1), 114–32. https://​doi​-org​.pros2​.lib​.unimi​.it/​10​ .1111/​1748​-8583​.12258. Dulebohn, J.H. and Marler, J.H. (2005). E-compensation: the potential to transform practice. In H.G. Gueutal and D.L. Stone (eds), The Brave New World of eHR: Human Resources Management in the Digital Age (pp.166–89). Jossey-Bass. Galanaki, E., Lazazzara, A., and Parry, E. (2019). A cross-national analysis of E-HRM configurations: integrating the information technology and HRM perspectives. In A. Lazazzara, R. Nacamulli, C. Rossignoli, and S. Za (eds), Organizing for Digital Innovation. Lecture Notes in Information Systems and Organization (pp. 261–76). https://​doi​.org/​10​.1007/​978​ -3​-319​-90500​-6​_2. García-Arroyo, J.A., Segovia, A.O., and Peiró, J.M. (2019). Meta-analytical review of teacher burnout across 36 societies: the role of national learning assessments and gender egalitarianism. Psychology & Health, 34(6), 733–53. https://​doi​.org/​10​.1080/​08870446​.2019​ .1568013 Góes, A. and De Oliveira, R.C.L. (2020). A process for human resource performance evaluation using computational intelligence: an approach using a combination of rule-based classifiers and supervised learning algorithms. IEEE Access, 8, 39403–19. https://​doi​-org​ .pros2​.lib​.unimi​.it/​10​.1109/​ACCESS​.2020​.2975485. Huai, C. and Chai, L. (2016). A Bibliometric Analysis on the Performance and Underlying Dynamic Patterns of Water Security Research. Scientometrics. Kehoe, J.F., Dickter, D.N., Russell, D.P., and Sacco, J.M. (2005). E-selection. In H.G. Gueutal and D.L. Stone (eds), The Brave New World of eHR: Human Resource Management in the Digital Age. Jossey-Bass.

150  Research handbook on human resource management and disruptive technologies

Keupp, M.M., Palmié, M., and Gassmann, O. (2012). The strategic management of innovation: a systematic review and paths for future research. International Journal of Management Reviews, 14(4), 367–90. Lamboglia, R., Lavorato, D., Scornavacca, E., and Za, S. (2020). Exploring the relationship between audit and technology. A bibliometric analysis. Meditari Accountancy Research, 29(5), 1233–60. https://​doi​.org/​10​.1108/​MEDAR​-03​-2020​-0836. Mahmoud, A.A., Shawabkeh, T.A., Salameh, W.A., and Al Amro, I. (2019). Performance predicting in hiring process and performance appraisals using machine learning. 10th International Conference on Information and Communication Systems (ICICS). IEEE, pp.  110–15. https://​doi​-org​.pros2​.lib​.unimi​.it/​10​.1109/​IACS​.2019​.8809154. Matt, D.T., Orzes, G., Rauch, E., and Dallasega, P. (2020). Urban production – a socially sustainable factory concept to overcome shortcomings of qualified workers in smart SMEs. Computers and Industrial Engineering, 139. https://​doi​.org/​10​.1016/​j. cie.2018.08.035. Nankervis, A., Connell, J., Cameron, R., Montague, A., and Prikshat, V. (2019). Are we there yet? Australian HR professionals and the Fourth Industrial Revolution. Asia Pacific Journal of Human Resources. https://​doi​.org/​10​.1111/​1744​-7941​.12245. Nørskov, S., Damholdt, M.F., Ulhøi, J.P., Jensen, M.B., Ess, C., and Seibt, J. (2020). Applicant fairness perceptions of a robot-mediated job interview: a video vignette-based experimental survey. Frontiers in Robotics and AI (November). https://​doi​.org/​10​.3389/​ frobt​.2020​.586263. O’Donovan, D. (2019). HRM in the organization: an overview. In C. Machado and J. Davim (eds), Management Science Management and Industrial Engineering (pp. 75–110). Springer. Office for AI (2019). Understanding Artificial Intelligence. Retrieved 1 December 2021. https://​www​.gov​.uk/​government/​publications/​understanding​-artificial​-intelligence. Sitzmann, T. and Weinhardt, J. M. (2019). Approaching evaluation from a multilevel perspective: a comprehensive analysis of the indicators of training effectiveness. Human Resource Management Review, 29(2), 253–69. Sivathanu, B. and Pillai, R. (2018). Smart HR 4.0 – how industry 4.0 is disrupting HR. Human Resource Management International Digest, 26(4), 7–11. https://​doi​.org/​10​.1108/​HRMID​ -04​-2018​-0059. Sony, M. and Naik, S. (2019). Key ingredients for evaluating Industry 4.0 readiness for organizations: a literature review. Benchmarking. https://​doi​.org/​10​.1108/​BIJ​-09​-2018​-0284. Stone, D.L., Deadrick, D.L., Lukaszewski, K.M., and Johnson, R.D. (2015). The influence of technology on the future of human resource management. Human Resource Management Review, 25(2), 216–31. Strohmeier, S. (2020). Digital human resource management: a conceptual clarification. German Journal of Human Resource Management, 34(3), 345–65. https://​doi​.org/​10​.1177/​ 2397002220921131. Suen, H.Y., Chen, M.Y.C., and Lu, S.H. (2019). Does the use of synchrony and artificial intelligence in video interviews affect interview ratings and applicant attitudes? Computers in Human Behavior, 98, 93–101. https://​doi​-org​.pros2​.lib​.unimi​.it/​10​.1016/​j​.chb​.2019​.04​.012. Tranfield, D., Denyer, D., and Smart, P. (2003). Towards a methodology for developing evidence – informed management knowledge by means of systematic review. British Journal of Management, 14(3), 207–22. Upadhyay, A.K. and Khandelwal, K. (2018). Applying artificial intelligence: implications for recruitment. Strategic HR Review, 17(5), 255–8. https://​doi​.org/​10​.1108/​SHR​-07​-2018​ -0051. Van Esch, P. J., Black, S., and Ferolie, J. (2019). Marketing AI recruitment: The next phase in job application and selection. Computers in Human Behavior, 90, 215–22. https://​doi​.org/​ 10​.1016/​j​.chb​.2018​.09​.009.

Is artificial intelligence disrupting human resource management?  151

Vrontis, D., Christofi, M., Pereira, V., Tarba, S., Makrides, A., and Trichina, E. (2021). Artificial intelligence, robotics, advanced technologies and human resource management: a systematic review. The International Journal of Human Resource Management, 1–30. https://​doi​-org​.pros1​.lib​.unimi​.it/​10​.1080/​09585192​.2020​.1871398. Wu, Z.X., Nkambou, R., and Bourdeau, J. (2012). Cultural intelligence decision support system for business activities. The Second International Conference on Business Intelligence and Technology, BUSTECH. Za, S., Spagnoletti, P., Winter, R., and Mettler, T. (2018). Exploring foundations for using simulations in IS research. Communications of the Association for Information Systems, 42, 268–300. Zehir, C., Karaboğa, T., and Başar, D. (2020). The Transformation of Human Resource Management and Its Impact on Overall Business Performance: Big Data Analytics and AI Technologies in Strategic HRM. Digital Business Strategies in Blockchain Ecosystems, 265–79. Zupic, I. and Cater, T. (2015). Bibliometric methods in management and organization. Organizational Research Methods, 18(3), 429–72.

11. Engaging intentionally disconnected workers: what can HR managers in facilities with workplace personal technology bans do? Melina Bumann and Michael Wasserman

INTRODUCTION Engagement is an important construct for both workers and their managers. But an important open question is: can HR managers use digital technology to improve worker engagement, especially among workers that are difficult to connect with during regular working hours, using typical web-based tools? This chapter looks at this specific type of worker – those in operational settings where the use of personal technology such as mobile phones and tablets is forbidden, including many manufacturing plants, distribution centers or facilities deemed as secure for information security or intellectual property protection. Sometimes this restriction is because of worker safety, sometimes to secure against corporate espionage and sometimes because of negotiated work rules designed to improve productivity. Regardless of the rationale, these workers are intentionally cut off from specific tools that HR managers often use to improve engagement in facilities where workers have regular access to web-enabled PC workstations and/or personal technology such as mobile phones and tablets. This ban on personal technology results in the elimination of what some companies have found to be useful tools for worker engagement, such as worker surveys, information on opportunities for job enhancement such as training, news and information about the company and the ability to connect with other employees in real time. Since engaged employees have been shown to be more productive and tend to stay longer in the organization, perhaps the use of technologies in the workplace can serve a similar function as a worker’s personal technology, and a technological solution might be a positive intervention in workspaces where personal technologies are forbidden. For example, simple alternative technologies can be placed in spaces where workers can access them either during break time (in a lunchroom or canteen, for example) or in a sanctioned way during work hours, for example during periods of machine maintenance, in areas near where other information such as Kanban boards or traditional bulletin boards are located. In this chapter, we will explore this disruptive strategy of redefining how, when and where HR managers can use tools to improve engagement for these intentionally disconnected workers. In an exploratory study, we test this disruptive HR strategy 152

Engaging intentionally disconnected workers  153

by moving from limited face-to-face and paper-based engagement activities to interactive digital engagement strategies for a group of these disconnected workers. We do this by developing and testing a simple digital platform. Our results lead to a discussion about the potential for a simple digital technology to serve as a conduit for further disruptive HR technologies such as chatbots and HR analytics that are related to engagement. We will then discuss our results, which will lead us to recommend that simple digital solutions aimed at this underserved segment of the workforce can have positive, disruptive effects for multiple stakeholders.

WORKFORCE ENGAGEMENT: A BRIEF OVERVIEW Workforce engagement is defined as the participation of organizational members in their work responsibilities, self-engagement and self-expression at physical, cognitive and emotional levels in their work lives (Kahn, 1990). The engagement construct enriched the toolkit HR managers could use with workers to improve performance. Before HR researchers and managers developed understanding and use of worker engagement, employee satisfaction was often a focus of HR managers. However, satisfaction did not have a strong empirical link to individual performance. Furthermore, satisfaction was more of a one-way construct, relating more to the worker than the company or even the interaction between worker and organization. This shift from satisfaction to engagement helped HR managers work strategically to improve a variety of other positive individual outcomes, including commitment, mental alertness, energy and greater intensity of interpersonal connections (Maslach et al., 2001; Schaufeli and Salanova, 2007; Macey and Schneider, 2008; Leiter and Bakker, 2010; Chandel, 2018; Sun, 2019). It is important to note that engagement starts with personal work experiences and individual decisions that cannot be forced. Engagement has both psychological and behavioral components for individual workers, including drive, passion and focused effort (Macey and Schneider, 2008). There are also positive impacts of engagement on organizational performance. Heintzman and Marson (2005) asserted that the concept of workforce engagement originated from the literature on social exchange theory and separated organizational commitment and work engagement into different types of engagement. Engagement has been linked to favorable organizational outcomes such as employee retention, productivity, profitability, customer loyalty and safety (Vance, 2016). Baumruk and Gorman (2006) found that employees who are engaged consistently exhibit three general behaviors that boost organizational performance: promotion and referral of the company to others, retention intentions and a willingness to exceed expected investments in time, effort, and initiative to contribute to the company’s success. Both academic research and the popular business press suggest that companies with engaged workers grow revenues faster than the industry average (Sorensen, 2013) and have higher operating and net profit margins than companies with a less engaged workforce (Borysenko, 2019). Worker engagement has been associated with a range of outcomes including attendance, employee turnover, productivity, innova-

154  Research handbook on human resource management and disruptive technologies

tiveness and higher error rates (Hakanen et al., 2008; Robertson and Cooper, 2010; Taris et al., 2010; Christian et al., 2011). Finally, if profitability is positively related to engagement, it is feasible to argue that shareholders can also share in the benefits of employee engagement. This logic leads to the question of what is the role of HR managers in this context.

HOW CAN HR MANAGERS IMPROVE WORKER ENGAGEMENT? Many strategies have been offered in the literature to help HR managers promote worker engagement. These strategies can be grouped into three categories: communication, recognition and accountability. The first two are obvious. Communication is a well-known strategy that HR managers need to use to build engagement (Woodruffe, 2006; Kontakos and Stepp, 2007; Swarnalatha and Prasanna, 2012). Recognition is important as workers who are more engaged in their work want to receive both monetary and nonmonetary benefits, including praise and recognition (Brun and Dugas, 2008). The third category, accountability, and its related construct of autonomy is often more complex to implement. Giving employees more autonomy encourages their independent thinking and empowers them to choose the best way to complete their tasks, as long as they deliver the desired results. That means managing for results, rather than trying to manage all the steps along the way (Rapp et al., 2006). Additionally, HR managers should be measuring and evaluating worker engagement as an essential aspect of performance management (Gruman and Saks, 2011). This process starts with regular assessment of worker engagement behavior in addition to work performance. However, understanding workforce engagement goes beyond tracking. The feedback must be translated into action, including clear and consistent communication of expectations and, where possible, sharing authority with workers through participatory decision-making so that they experience a sense of belonging and are more engaged in achieving them (Lartey, 2021). You may be reading this and thinking how obvious all of this is, and questioning whether you should finish reading this chapter. But hold on! As mentioned earlier, there is a large, important and often underserved group of workers that, because of workplace policies, are difficult for HR managers to reach with all three of the strategies mentioned above. These are workers who are not allowed to use personal technology such as mobile phones, tablets or internet-enabled personal computers. This group, who work in manufacturing plants or logistics facilities, are hard to reach with typical engagement strategies that are web-based and use either personal devices, such as phones or tablets, or typical personal computers. This access is taken for granted by those who work in typical offices, work remotely from a home office or in locations where use of mobile phones or tablets is allowed.

Engaging intentionally disconnected workers  155

Leave Your Device at the Door: Challenges in Implementing Engagement Strategies Many tools and strategies that HR managers use are designed for a desktop workstation with an internet-enabled personal computer. Some are designed to be used on mobile phones. In a manufacturing plant or fulfillment facility, employees do not have such a workstations and often personal devices such as mobile phones and tablets are not allowed. These rules exist for a variety of practical reasons. For example, in many countries, there are laws and regulations concerning health and safety in the workplace, including hygiene regulations for production facilities where food is handled (see, for example, Regulation (EC) No 852/2004). Companies in these industries require protective clothing, but also might ban personal electronics, as these can introduce pathogens into a clean workspace. Companies like Amazon also have a smartphone ban in their warehouses, not for hygiene reasons, but safety reasons. Samsung banned camera phones as early as 2003 to prevent the disclosure of their intellectual property. Whether it is for data security, protection of research and new product or process development, many companies do not want to risk proprietary data leaking to competitors, intentionally or unintentionally. Competitors can conduct industrial espionage by hacking into the camera or microphone of a worker’s smartphone. Additionally, companies that manufacture or ship personal electronic items have found it necessary to prevent employee theft (Fong, 2020). These workplace personal technology bans make it difficult for HR managers to communicate with workers via a digital channel in ways analogous to how they interact with employees that work in a traditional office setting and are allowed to use personal digital devices like mobile phones. This means HR managers cannot use these important and easy-to-use channels for strengthening employee engagement. Specifically, these workplace bans make it difficult for HR managers to implement all three types of strategies discussed earlier in this chapter: communication, recognition, and accountability. Without easy access via smartphones or web-based PC workstations, HR managers rely on physical bulletin boards, video monitors, and traditional face-to-face meetings to interact with these workers. As companies face worker shortages in manufacturing and distribution facilities, any gaps in opportunities to improve engagement seem essential to address. What are some potential ways to implement this disruptive strategy of improving engagement for intentionally disconnected workers? Perhaps HR managers can use digital tools. The idea of digital kiosks in break rooms or on shop floors is not new, but these types of technologies have not been well studied in terms of their impact on worker engagement. Are these kiosks or tablets disruptive in and of themselves? Probably not, but as we discuss, these platforms offer an open door into introducing other digital tools that are disruptive technologies that are likely important for these disconnected employees.

156  Research handbook on human resource management and disruptive technologies

The Disruptive Approach: Strategy Followed by a Platform for Future Disruptions First, let’s define what a disruptive technology is and is not. Christensen (1997) defined disruptive technologies as having lower cost and performance as measured by traditional measures. Utterback and Acee (2005) explained that disruptive technologies can be an important driver of enhancing access to new or underserved customers or stakeholders and can provide new functionality or benefits. Looking at disruption through this lens, we can see that the potential here for a disruptive solution is theoretically feasible. This disruption, using simple tools such as a kiosk or tablet, can enhance access and benefits to underserved stakeholders – in this case, intentionally disconnected workers. And importantly, although many take a Schumpeterian view of disruption – one of creative destruction, Utterback and Acee (2005) explain that disruption can be something positive for multiple stakeholders – no destruction is required. And we think, in this context, we are looking at a disruption that provides positive benefits to multiple stakeholders. It is a major change in how stakeholders interact. It changes some of the rules of engagement (e.g. how HR managers interact with intentionally disconnected workers), but the end results is a better workplace for everyone, with shared benefits for HR managers, workers and shareholders. Second, we should take a closer look at the nature of this disruption – where is the disturbance or disruption in the HR ecosystem? What is being disturbed or disrupted? Tekic and Koroteev (2019) provide a lens to look at the nature of the HR disturbance or disruption in this context of engagement and intentionally disconnected workers. Engaging these disconnected workers is potentially a substantial change to the value proposition. In this case, a shift in value provided by HR managers can occur. There is a ‘what’ – enhanced engagement opportunities, a ‘to whom’ – intentionally disconnected workers, a ‘how’ – using digital technologies like tablets or kiosks, a ‘where’ – in breakrooms or other locations onsite in facilities that ban personal digital devices, and a ‘when’ – workers access the tools during breaks and low-demand periods of the work day. The disruption disturbs the prior practices of face-to-face communication between HR managers and these workers, or one-way communication mechanisms such as bulletin boards, leaflets or video screens, or simply, the existence of a gap between HR managers and these workers when it comes to engagement. Third, using the Tekic and Koroteev (2019) lens, these engagement behaviors are now clearly critical for the success of many companies, given that, in many parts of the world, vacancies in key positions in logistics and manufacturing centers are now difficult to fill. These kiosk-based solutions we discuss are low-risk with low consequences for failure. And finally, given that these are mostly commodity products with limited need for customization through physical engineering or programming, these solutions can be used in a fail fast, fail cheap, lean approach. Are these truly disruptive technologies that we are discussing? Disruptive HR technologies can be looked at in many ways – from cutting edge tools such as artificial intelligence (Fisher and Howardson, 2022) to sensors/internet of things (Strohmeier, 2022) and big data (Marler and Boudreau, 2017); but also, disruption

Engaging intentionally disconnected workers  157

can exist where HR managers use digital tools to change the work that they do (see, for example, Bondarouk and Ruël, 2013 and Meijerink et al., 2018). To accomplish strategic HR change Strohmeier (2020) and Tekic and Koroteev (2019) argue that some degree of disruption is typically involved, and it is likely that disruption is a two-stage process. The first is a strategic disruption. Here is a group of workers that, in the past, was neither accessible through digital channels (intentionally disconnected workers) nor perceived to be of high value to the organization. But these workers are now more important in a post-COVID world (Afonso et al., 2021; Qin et al., 2021). HR managers need to develop a competitive strategy that considers the new, higher value, of these workers in logistic centers and factories and address the limitations these workers face by not having access to personal digital devices in the workplace. And to do this, HR managers must enhance the level of digital accessibility for disconnected workers to improve worker engagement. This attention by HR managers is even more critical as new disruptions to supply chains and traditional employment models, including COVID, the war in Ukraine and rapid climate change pile up. HR managers need to radically rethink engagement with many categories of employees, as multiple disruptions to business models and HR systems have made attracting and retaining workers in areas like logistics and factory work more critical. Especially, as work-from-home and phenomena like the great resignation create havoc with traditional HR recruiting, staffing and retention models. Figure 11.1 displays this disruptive strategic shift. This is a digital HRM disruption that, following the conceptual logic of Strohmeier (2020), is a process that aligns digital technologies to existing HR strategies (in this context, worker engagement). Applying Figure 11.1, HR creates a completely new HR strategy, formulated and implemented using digital technologies to create value by moving from the current state of little engagement of these works to a new state of intensive digital engagement. In this case, HR managers are engaging intentionally disconnected workers and creating value for workers, the HR function and shareholders by exploiting previously unexplored digital opportunities. In the next section, we present a conceptual model that takes these disruptions and tries to align research and practice to propose and test real solutions.

A CONCEPTUAL MODEL OF DIGITAL ENGAGEMENT FOR DISCONNECTED WORKERS Theory and existing research suggest that digital workplace tools such as kiosks could be used to enhance worker engagement. Integrating existing theory and research suggests that a process might look like Figure 11.2. The effectiveness of traditional employee engagement strategies focusing on communication, recognition and accountability are attenuated for those employees without personal digital access in the workplace. Although traditional engagement strategies can work for these workers using paper-based and face-to-face communication, using specifically designed digital workplace tools can serve as a disruptive technology in the work-

158  Research handbook on human resource management and disruptive technologies

Figure 11.1

Illustration of disruptive HR strategy of investing in worker engagement platforms for intentionally disconnected workers

place to help HR managers greatly enhance worker engagement among these important segments of workers, thus moderating the relationship between engagement strategies and worker engagement.

Figure 11.2

A model of using digital workplace tools to reach intentionally disconnected workers

Engaging intentionally disconnected workers  159

Designing Effective Digital Tools for Intentionally Disconnected Workers There are two questions to address when discussing the design of digital tools for this specific context. The first is accessibility. What type of device and where should it be located in order to be convenient and useful to the worker while addresses the concerns of the company, whether this be protection of intellectual property, health and safety or other reasons? In a factory or distribution setting, this needs to be in an area that is accessible to the worker, but neither an impediment to safety nor a distraction. Spaces that meet these criteria are limited, but might include breakrooms or lunchrooms, rather than entry ways or other high-traffic, potentially congested areas. Of course, privacy to read and respond to messages or polls should be considered as important. Observations of these spaces also suggest that lighting, noise, vibrations and other unique physical attributes of industrial settings should also be considered when choosing both the types of digital devices and the location of these devices in order to maximize the likelihood that employees will be willing and able to use these digital tools. Second, what type, design and form should the information take? We look here to the research on information design. Horn (1999) defined information design as providing the right information in the right format at the right time to the right people. According to Horn, information design should be intentional, so that: (1) Information can be quickly and accurately understood; (2) It allows people to navigate easily throughout the digital space; (3) Interacting with information is simple, natural and enjoyable; and (4) Information is easy to translate into effective action when action is desired. Whitehouse (1999) noted that individual needs should be thoughtfully considered when designing information to be presented. In this case, since engagement is a two-way process, both the needs of the intentionally disconnected worker, as well as needs of the HR managers, should be considered carefully. These guidelines can help HR managers design the information presented to workers for the three specific tasks discussed earlier: communicating (news and information to the worker and collecting feedback or ideas from the worker), recognition of the worker (personal communication from supervisors or peers, or badges or icons denoting accomplishments) and accountability (performance data, especially in comparison to goals). Even more specifically, information design includes typography, color and layout, especially the relationship between words and images (Redish, 2000). This goes beyond just aesthetics and should be carefully designed and tested by HR managers working with information designers to maximize the impact on the drivers of engagement. This includes the choice of content, the way the content appears and the ways users interact. The goal is to get workers to use and interact with relevant information to improve engagement, so specific design elements need to be considered (Shedroff, 1999). Specific examples of relevant information design in this context include

160  Research handbook on human resource management and disruptive technologies

human-computer interaction (HCI), which operates at the intersection between computer science and technology, as well as psychology and social sciences (Carroll, 1997). HCI focuses on improving the usability of computer systems and applications (Carroll, 1997) usually through designing graphical user interfaces that consider the environment, the user’s surroundings, the size and placement of the device and the user’s interaction with the device (Baer and Vacarra, 2008). Techniques such as interface design, modeling, data analysis and information retrieval are well supported by the HCI perspective both for consumers (Martinez-Toro et al., 2019) and in the workplace (Canedo et al., 2017).

SOME EXPLORATORY DATA COLLECTION Perhaps we can use technology to disrupt the typical relationship between HR managers and digitally disconnected workers to try and obtain benefits related to engagement for both parties. To learn more about the nature of this issue, the second author conducted a small set of exploratory interviews with HR managers that serve significant numbers of intentionally disconnected workers at two companies, one in the US and one in Europe. We asked questions about engagement efforts with workers in settings such as manufacturing lines and distribution centers where employees had no access to personal technology and little access to web-enabled PC workstations. We heard that HR managers do feel that there are challenges and frustrations to engaging workers in these settings and typically use face-to-face meetings and either video or physical boards to share information and connect with workers during working hours. HR managers saw this as a digital divide between white collar and blue collar workers and that there were problems in engaging intentionally disconnected workers, although neither company had any solid data on the extent of the issue. We took what we learned from the interviews with HR managers and used this opportunity to take two next steps. First, we created some prototypes to try to capture some of the engagement strategies and tools that HR managers thought might be especially challenging for those workers that were intentionally digitally disconnected to access. Second, we used the information to create a set of exploratory interviews with these intentionally disconnected workers. We used principles of information design and created prototypes that applied the basic theories of worker engagement mentioned above, including company/facility news, individual/team/facility performance information, individual feedback polls and recognition of teams and individuals. The prototypes were running on tablets and had all basic functionality in place. After we built the prototypes, we conducted a set of exploratory interviews with intentionally disconnected workers at fulfillment centers run by one of the two companies. We even created a few prototypes of workplace engagement using information design principles to do a basic test of the viability of using a simple yet potentially disruptive idea: company-owned, interactive kiosks or tablets securely mounted on a table (like one might see at an airport) with digital engagement tools

Engaging intentionally disconnected workers  161

provided in worker break rooms in facilities where workers were forbidden from bringing personal devices into the workplace. We conducted a set of 12 structured interviews of intentionally disconnected employees to evaluate their level of engagement, exploring the three strategies for employee engagement (communication, recognition and accountability), as well as four state engagement variables identified by Macey and Schneider (2008), satisfaction, involvement, commitment and empowerment. Afterward, interviewees were asked to explore a prototype of a digital tool designed to be a two-way tool for communication, recognition and accountability. Interviewees were asked to navigate each of the key functions and asked about their willingness to increase their engagement if the prototype was introduced into their work environment. This was only intended to be an exploratory set of interviews, and this was not a scientifically random sample. In fact, all interviewees were volunteers, and it is likely that volunteers are already exhibiting engagement behaviors. In fact, all of the interviewees described existing engagement emotions and behaviors and discussed engagement with the company in a generally positive way. However, there were general perceptions that the company could do more to support engagement with workers. We were interested to learn that these intentionally disconnected employees showed a general willingness to use a digital interface to the company such as a kiosk or a table-mounted tablet during breaks or other downtime in the facility. Nine of twelve participants made it clear that they would definitively use a tool like this if it was implemented at the company. For some participants, the prototype was intuitively understandable at first glance. Participants were able to navigate without problems and understood the purpose of the information provided. More specifically, most of the participants believed that the digitalization of the feedback process would be beneficial. Most pointed out that it would be a simple and useful tool to increase worker engagement by reducing barriers to providing feedback to managers. Similarly, most participants commented positively on the survey function of the prototype. Interestingly, most participants reported that they found the available information about company news, e.g. regarding company events, sufficient. They mentioned pamphlets and posters as well as monitors as examples. However, many participants preferred the format of interacting with a touchscreen and the comfort of reading and researching at their own speed. They also pointed out that it would be easier to read up only on news that interested them. Combining those statements leads to the exploratory finding that providing intentionally disconnected workers with some easily available digital tools, designed for engagement, appears to be a strategy worth more testing by organizations that employ intentionally disconnected workers. Specifically, participants were most interested in news and feedback, and least interested in a performance overview or an employee of the month award. The collection of findings implies that a digital interface for interaction between employees and the company has the potential to increase employee engagement. Particularly in the area of involvement and empowerment, the greatest improvements are possible, as the participants responded especially

162  Research handbook on human resource management and disruptive technologies

to the functions from the areas of recognition and communication. Thus, it can be concluded that a digital interface for interaction between companies and employees may be an effective tool for strengthening workforce engagement in the factory or fulfilment context. In summary, initial evidence found that a simple but disruptive tool, a kiosk or tablet in a workspace might make it easier for HR managers to improve engagement among this group of workers using straightforward communication and recognition strategies. Although this is only one company and a small set of workers, our findings suggest that companies should be exploring whether this option might work for their specific situation.

CONCLUSION AND RECOMMENDATIONS FOR ACTION Additional, broader research should be conducted on engagement issues with intentionally disconnected workers, especially those in factories or distribution centers where HR managers are struggling to retain and hire new workers. If these straightforward and inexpensive solutions can improve engagement, and thus retention, it can be a win-win-win for top managers, HR managers and workers. Ideas for additional content, as voiced by the employees, include monthly updates on corporate goals and directions, idea collection schemes linked to rewards for idea generation and community outreach activities such as volunteering for social activities. While the above ideas are straightforward, we recommend that a digital platform can also be used as a mechanism to launch more disruptive technologies related to engagement, such as natural language processing-based chatbots (Smith, 2019; Miklosik et al., 2021), accessibility of smart HR analytics (Malik et al., 2022), and AI-driven training opportunities (Maity, 2019; Malik et al., 2022). With such a platform, a wide variety of individualized employee experiences can be tied to worker engagement (Malik et al., 2020) for this group of intentionally disconnected employees. We conclude by arguing that this overlooked segment of workers deserves a disruption in the way they interact with HR managers, disruption that should result in more aligned individual and organization performance. We think, in this situation, our recommendations result both in some inexpensive, fast and simple actions and provide a longer-term conduit that HR managers can use to improve individual and organizational performance. Both the short- and long-run elements of our proposed solution aim to improve a challenging and important HR context – engaging and retaining critical workers in manufacturing and distribution facilities. Ultimately, we found that specific digital solutions aimed at intentionally disconnected workers – an important and underserved segment of the workforce – can have disruptive, yet positive effects for these disconnected workers, their managers and the shareholders of the companies for whom they work.

Engaging intentionally disconnected workers  163

REFERENCES Afonso, T., Alves, A. C., and Carneiro, P. (2021). Lean thinking, logistic and ergonomics: synergetic triad to prepare shop floor work systems to face pandemic situations. International Journal of Global Business and Competitiveness, 16(1), 62–76. Baer, K. and Vacarra, J. (2008). Information Design Workbook: Graphic Approaches, Solutions, and Inspiration. Rockport Publishers. Baumruk, R. and Gorman, B. (2006). Why managers are crucial to increasing performance. Strategic HR Review, 5(2), 24–7. Bondarouk, T. and Ruël, H. (2013). The strategic value of e-HRM: results from an exploratory study in a governmental organization. International Journal of Human Resource Management, 24(2), 391–414. Borysenko, K. (2019). How much are your disengaged employees costing you? Forbes, May 2. https://​www​.forbes​.com/​sites/​karlynborysenko/​2019/​05/​02/​how​-much​-are​-your​ -disengaged​-employees​-costing​-you. Brun, J. P. and Dugas, N. (2008). An analysis of employee recognition: perspectives on human resources practices. The International Journal of Human Resource Management, 19(4), 716–30. Canedo, J., Graen, G., Grace, M., and Johnson, R. (2017). Navigating the new workplace: technology, millennials, and accelerating HR innovation. AIS Transactions on Human-Computer Interaction, 9, 243–60. https://​doi​.org/​10​.17705/​1thci​.00097. Carroll, J. M. (1997). Human-computer interaction: psychology as a science of design. International Journal of Human-Computer Studies, 46(4), 501–22. Chandel, P. (2018). The evolution of employee engagement, a unique construct. International Journal of Human Resource Management and Research, 8(6), 199–216. Christensen, C. M. (1997). The Innovator’s Dilemma: When New Technologies Cause Great Firms to Fail. Harvard Business Review Press. Christian, M. S., Garza, A. S., and Slaughter, J. E. (2011). Work engagement: a quantitative review and test of its relations with task and contextual performance. Personnel Psychology, 64(1), 89–136. Fisher, S. L. and Howardson, G. N. (2022). Fairness of artificial intelligence in human resources-held to a higher standard? In S. Strohmeier (ed.), Handbook of Research on Artificial Intelligence in Human Resource Management, Edward Elgar, 303–22. Fong, M. (2020). Employee smartphones: to ban or not to ban? Forbes, February 25. https://​ www​.forbes​.com/​sites/​forbestechcouncil/​2020/​02/​25/​employee​-smartphones​-to​-ban​-or​ -not​-to​-ban. Gruman, J. A. and Saks, A. M. (2011). Performance management and employee engagement. Human Resource Management Review, 21(2), 123–36. https://​doi​.org/​10​.1016/​j​.hrmr​.2010​ .09​.004. Hakanen, J. J., Perhoniemi, R., and Toppinen-Tanner, S. (2008). Positive gain spirals at work: from job resources to work engagement, personal initiative and work-unit innovativeness. Journal of Vocational Behavior, 73(1), 78–91. Heintzman, R. and Marson, B. (2005). People, service and trust: is there a public sector service value chain? International Review of Administrative Sciences, 71(4), 549–75. https://​ journals​.sagepub​.com/​doi/​pdf/​10​.1177/​0020852305059599. Horn, R. E. (1999). Information design: emergence of a new profession. In R. E. Jacobsen (ed.), Information Design, MIT Press, 15–33. Kahn, W. A. (1990). Psychological conditions of personal engagement and disengagement at work. Academy of Management Journal, 33(4), 692–724. https://​doi​.org/​10​.2307/​256287. Kontakos, A. and Stepp, P. (2007). Employee Engagement and Fairness in the Workplace. Center for Advanced Human Resource Studies, Cornell University.

164  Research handbook on human resource management and disruptive technologies

Lartey, F. M. (2021). Impact of career planning, employee autonomy, and manager recognition on employee engagement. Journal of Human Resource and Sustainability Studies, 9(02), 135–58. https://​doi​.org/​10​.4236/​jhrss​.2021​.92010. Leiter, M. P. and Bakker, A. B. (2010). Work engagement: introduction. In A. B. Bakker and M. P. Leiter (eds), Work Engagement: A Handbook of Essential Theory and Research, Psychology Press, 1–9. Macey, W. H. and Schneider, B. (2008). The meaning of employee engagement. Industrial and Organizational Psychology, 1(1), 3–30. Maity, S. (2019). Identifying opportunities for artificial intelligence in the evolution of training and development practices. Journal of Management Development, 38(8), 651–63. Malik, A., Budhwar, P., Patel, C., and Srikanth, N. R. (2020). May the bots be with you! Delivering HR cost-effectiveness and individualised employee experiences in an MNE. International Journal of Human Resource Management, 1–31. Malik, A., Thevisuthan, P., and De Sliva, T. (2022). Artificial intelligence, employee engagement, experience, and HRM. In A. Malik (ed.), Strategic Human Resource Management and Employment Relations, Springer, 171–84. Marler, J. H. and Boudreau, J. W. (2017). An evidence-based review of HR analytics. International Journal of Human Resource Management, 28(1), 3–26. Martinez-Toro, G. M., Ariza-Zabala, G. C., Bautista, D. R., and Romero-Riaño, E. (2019). Human-computer interaction in transport, a systematic literature review. Journal of Physics: Conference Series (Vol. 1409, No. 1, p. 012002, November). IOP Publishing. Maslach, C., Schaufeli, W. B., and Leiter, M. P. (2001). Job burnout. Annual Review of Psychology, 52, 397–422. Meijerink, J., Boons, M., Keegan, A., and Marler, J. (2018). Special issue of the international journal of human resource management: digitization and the transformation of human resource management. The International Journal of Human Resource Management, 1–6. Miklosik, A., Evans, N., and Qureshi, A. M. A. (2021). The use of chatbots in digital business transformation: a systematic literature review. IEEE Access, 9. https://​doi​.org/​10​.1109/​ ACCESS​.2021​.3100885. Qin, X., Godil, D. I., Khan, M. K., Sarwat, S., Alam, S., and Janjua, L. (2021). Investigating the effects of COVID-19 and public health expenditure on global supply chain operations: an empirical study. Operations Management Research, 1–13. Rapp, A., Ahearne, M., Mathieu, J., and Schillewaert, N. (2006). The impact of knowledge and empowerment on working smart and working hard: the moderating role of experience. International Journal of Research in Marketing, 23(3), 279–93. https://​doi​.org/​10​.1016/​j​ .ijresmar​.2006​.02​.003. Redish, J. C. G. (2000). What is information design? Technical Communication, 47(2), 163–6. Regulation (EC) No 852/2004 of the European Parliament and of the Council of 29 April 2004 on the hygiene of foodstuffs. Robertson, I. T. and Cooper, C. L. (2010). Full engagement: the integration of employee engagement and psychological well-being. Leadership & Organization Development Journal, 31(4), 324–36. Schaufeli, W. and Salanova, M. (2007). Work engagement: an emerging psychological concept and its implications for organizations. In S. W. Gilliland, D. D. Steiner, and D. P. Skarlicki (eds), Managing Social and Ethical Issues in Organizations, Information Age Publishing. https://​www​.wilmarschaufeli​.nl/​publications/​Schaufeli/​273​.pdf. Shedroff, N. (1999). Information interaction design: a unified field theory of design. In R. E. Jacobsen (ed.), Information Design, MIT Press, 267–93. Smith, C. (2019). An employee’s best friend? How AI can boost employee engagement and performance. Strategic HR Review, 18(1), 17–20. Sorenson, S. (2013). How employee engagement drives growth. Gallup Business Journal, 1, 1–4.

Engaging intentionally disconnected workers  165

Strohmeier, S. (2020). Digital human resource management: a conceptual clarification. German Journal of Human Resource Management, 34(3), 345–65. Strohmeier, S. (2022). Handbook of Research on Artificial Intelligence in Human Resource Management. Edward Elgar. Sun, L. (2019). Employee engagement: a literature review. International Journal of Human Resource Studies, 9(1), 63–80. Swarnalatha, D. C. and Prasanna, T. S. (2012). Employee engagement: the key to organizational success. International Journal of Management, 3(3), 216–27. Taris, T. W., Schaufeli, W. B., and Shimazu, A. (2010). The push and pull of work: about the difference between workaholism and work engagement. In A. B. Bakker and M. P. Leite (eds), Work Engagement: A Handbook of Essential Theory and Research, Psychology Press, 39–53. Tekic, Z. and Koroteev, D. (2019). From disruptively digital to proudly analog: a holistic typology of digital transformation strategies. Business Horizons, 62(6), 683–93. Utterback, J. M. and Acee, H. J. (2005). Disruptive technologies: an expanded view. International Journal of Innovation Management, 9(1), 1–17. Vance, R. (2016). Employee Engagement and Commitment A Guide to Understanding, Measuring and Increasing Engagement in your Organization. Society of Human Resource Management. Accessed at https://​www​.shrm​.org. Whitehouse, R. (1999). The uniqueness of individual perception. In R. E. Jacobsen (ed.), Information Design, MIT Press, 103–29. Woodruffe, C. (2006). The crucial importance of employee engagement. Human Resource Management International Digest, 14(1), 3–5.

12. What decision-makers need to know about digitalised talent management Sharna Wiblen

INTRODUCTION Spoiler alert: All talent decisions are subjective. The ongoing concerns about acquiring and retaining talent show no signs of abating. News headlines declare a new(ish) talent crisis associated with the COVID-19 pandemic and the increasing pressure on talent supply chains. References to the ‘great resignation’ feature in public discourse, with organisations, yet again, told to compete for a finite supply of talent. The pandemic has also (again) illuminated the importance of ‘the workforce’ in business continuity: ‘the workforce’ is helping organisations ‘keep the lights on’ as organisations navigate their current externalities and seek to ‘come out the other side’. The pandemic reminds us of the need to consider how to organise and manage workforces for operational and strategic purposes. Talent management – notwithstanding crises – is fundamentally complex. Leaders are told to pay attention to their talent because talent is their greatest asset. More specifically, the workforce and talent are worthy of proactive management and investment. Significantly, talent is essential to operational needs and strategy execution because strategies are actualised (or not) because of the workforce’s actions (or inactions). The challenge, however, is that talent, as noted above, is also (potentially) an organisation’s most considerable expense (Davenport et al., 2010). Leaders are continually making decisions about whether to focus on the talent needs of today or some future day and whether a policy of practice is an investment or a required expense. Digitalised talent management is inherently complex, and it is only by recognising the entangled relationship between talent and technology that we can foster better talent decisions. My research and teaching have taught me that focusing on specifics can limit conversations and misdirect attention towards debating definitions rather than understanding the complexity associated with all forms of technology. I decided, therefore, to frame ‘digitalisation’ broadly. What follows is the presentation of different aspects of talent and digitalisation that decision-makers should be aware of when making talent decisions: ● Digitalised talent management is a topic worth paying attention to. ● Understand that talent management and technology are interrelated. ● Reflect on whether current digitalisation was a strategic choice or pandemic induced. 166

What decision-makers need to know about digitalised talent management  167

● Recognise that technology is only a tool. ● Understand that talent management and human resource management are not the same. ● Recognise that talent management focuses on judging value. ● Recognise that digitalisation frames talent as a static concept. ● Recognise that algorithms are not objective. ● Recognise that talent decisions are always subjective. I do not imply that these aspects represent everything you need to know, or that there is a set ‘recipe’ to follow. Reflecting on each element, however, assists in enhancing our understanding of the human-technology interface.

PAY ATTENTION TO DIGITALISED TALENT MANAGEMENT Industry and consultancy texts frequently note that talent management and digitalisation are at the top of priority lists (for example see PwC, 2017, 2019, 2020). While talk about the great resignation has (momentarily) superseded broader concerns about the impact of technological innovations, there is no mistaking that digitalisation continues to transform workplaces. Both hardware and software innovations provide new and revised working methods with extensive evidence that technology impacts how we work. Digitalisation raises many questions about human-machine work combinations and the benefits of automating tasks and occupations (Jesuthasan and Boudreau, 2018). Digitalisation, automation and technological ways of working replace specific tasks and jobs while simultaneously creating others. Many media outlets claim that technology is the enemy of some, with some headlines cautioning individual workers against technological innovations – aka Robots will come and take their jobs. Regardless of the external sentiment, media headlines and report findings, decision-makers must pay attention to digitalised talent management. Technology already plays a role in how many organisations organise, manage and undertake talent management practices. New challenges arise as more recent versions of certain technologies are released (think HR technology system update) and more unique innovations emerge (think about what comes after artificial intelligence and machine learning). Decision-making accompanies any new challenge as leaders decide where, when and how more contemporary forms of digitalisation will replace current ways of working and managing. Decision-makers must keep an eye on improving current practices and the potential influence of innovations on specific procedures, organisations, industries and society. Decision-makers, therefore, must pay attention to the ever-changing technology landscape. Decision-makers should dream about digitalised talent management’s potential good and sinister implications. Decision-makers, I implore you to keep

168  Research handbook on human resource management and disruptive technologies

your eyes open and make the deliberate choice to pay attention to digitalised talent management.

UNDERSTAND THAT TALENT MANAGEMENT AND TECHNOLOGY ARE INTERRELATED Talent management and digitalisation are inherently interrelated. Technology is used to manage workforces and shape workforce structures and compositions. Newer technological innovations result in new approaches to managing talent (Wiblen et al., 2010) and impact workers, jobs and careers (Jesuthasan and Boudreau, 2018). Inversely, talent availability influences whether organisations select and deploy specific technological innovations. Neither talent nor technology, therefore, is helpful in its own right. Effective talent management – the management of valuable individuals and groups of individuals – requires strategically aligned decisions and practices. Data, information and knowledge should inform talent-based choices and techniques. Data and information are captured, stored and analysed in conjunction with an information system – a pen and paper, Excel file, standalone or integrated human resource system (HRIS) or other technology. Digital transformation, Davenport and Redman (2020) state, requires talent in four key areas: technology, data, process and organisational change. A holistic perspective is vital because digitalised talent management occurs within broader systems. Appreciating both talent management and digitalisation domains is essential in devising project outcomes. Decision-makers can no longer treat technology as a separate and distinct phenomenon. Instead, technology is part of how we work (Orlikowski and Scott, 2008; Schatzki, 2010; Beane and Orlikowski, 2015). Technology assists with organising and managing talent. When talking about talent management, decision-makers should know that talent decisions include an element of technology. Decision-makers need to understand that the selection, implementation or revision of technology – whether a software or hardware component – has an associated talent element. Digitalisation, therefore, in all its forms, includes a human part. By recognising the intersectionality between talent and technology, we can make better decisions that intentionally recognise, not trivialise, the system-based nature of workforce practices.

WHETHER TODAY’S DIGITALISATION IS PANDEMIC INDUCED Let us then reflect on digitalisation within the current pandemic context. Some leaders embraced technological innovations and transitioned toward the digitalised workplace before March 2020. Steps towards increased digitalisation resulted from internal choices; because they wanted to. These leaders have had time to iterate and

What decision-makers need to know about digitalised talent management  169

amend internal policies and practices within the context of non-pandemic working methods. Other senior leaders have turned to technology to facilitate working methods because of the pandemic. These leaders reacted to the pandemic. These leaders may have modified practices to cater to the needs of remote workers and remote teams because they had to. As I ask for in the first section, some leaders were not paying attention to the ever-changing technology landscape and dreaming about the implications for digitalised talent management. The need or desire to transition towards digital working methods was minimal or non-existent. Increasing digitalisation was not previously deemed core to operational requirements or an intentional component of the organisation’s strategic ambitions. Instead, such organisations were forced into implementing digital ways of working. Decisions about technology platform adoption may have been reactive. Worst still is where technology decisions were driven by what they could get or afford at the time. Decision-makers can pause and reflect on whether today’s digitalisation resulted from the pandemic. Were leaders seeking to make a digital change before the pandemic with the pandemic accelerating the (already existing) implementation process? Or are their current ways of organising and managing the workforce an outcome of the pandemic? To what extent are digitalised practises the result of reacting to 2020–22 operational needs and business continuity? Answering yes, they are largely reactive, to the latter question potentially frames digitalisation as an immediate expense. In contrast, affirmative answers to the initial idea see digitalisation as the outcome of reflection and due process. Regardless of your financial statements, decision-makers benefit from knowing whether today’s digitalisation mirrors a strategic way. Suppose the digitalisation of today is not indicative of a strategic way. In that case, you need to re-evaluate the usefulness and relevance of the technology decision within what the organisation needs for your future day. Digitalisation choices made within the context of the operational needs of a pandemic may reflect requirements for a specific day. That does not mean that the digitalisation decision is the best for your strategy. You can revise the decision and make a new or different choice when framed within the context of what is needed to organise and manage your workforce in the future.

TECHNOLOGY IS A TOOL. DIGITALISATION IS NEVER THE DESTINATION Decision-makers know that technology helps organisations do what they need to do. Technology-enabled ways are more convenient than the manual and paper-based processes they replaced. Technological components are foundational to scaling working methods. Digitalisation assists with automating specific tasks and codifiable processes.

170  Research handbook on human resource management and disruptive technologies

Digitalisation enhances talent management processes. However, problems arise when framing technology as a panacea. Technology – be it digitalisation, artificial intelligence, machine learning, or others – cannot, in and of itself, make better talent decisions. In and of itself, technology cannot improve talent management. In and of itself, technology cannot make the world better – because technology is a tool. Digitalisation is not the destination. The digitalisation helps – but digitalisation is not the ‘where’ in ‘somewhere’. Decision-makers may ask – could technology help us get to our ‘where’? Technology can be part of the ‘how’ of pursuing strategic ambitions and goals. However, selecting, implementing, and using technology is not an end goal. Studies frequently show that the reality of technology is not a panacea, with many project goals largely unrealised. Theoretically and methodologically informed studies of technology implementations often report the inability to realise the benefits sprouted in the hype (for example, Grant et al., 2009; Dery et al., 2013; Wiblen and Marler, 2021). The reason is that technology is, as noted above, only a tool. Information technology has both physical and procedural dimensions. Physical components include the hardware, software and communication network infrastructures (Orlikowski and Scott, 2008). While these are separate from individuals, the physical aspects of technology are nothing without the humans who engage with it. Therefore, decision-makers, in recognising that technology is a tool not a destination, must reflect on humans’ role. Understanding the perceptions of humans who use the technology is crucial to understanding digitalisation’s value and limitations. (See Wiblen, 2016, to learn more about the factors that influence technology use in talent management; Wiblen et al., 2012; Wiblen and Marler, 2021.)

TALENT MANAGEMENT AND HUMAN RESOURCE MANAGEMENT ARE NOT THE SAME Decision-makers should know that talent management and human resource management (HRM) are not synonymous. Talent management and HRM are distinct. Talent management starts with strategy: strategy, strategy, strategy. Specifically, talent management focuses on identifying and developing individuals who play a more significant role in executing your team’s or organisation’s strategic ambitions and goals. Other ways to differentiate between talent management and human resource management are: ● HRM policies and practices apply to the entire workforce. These policies guide everyone who works for your organisation. However, talent management policies and practices apply to some categories of individuals within the workforce (think high performers, future leaders, etc.). Talent management only applies to talent – a select group/s of someones.

What decision-makers need to know about digitalised talent management  171

● HRM emphasises procedural and distributive justice whereby the desired outcomes are to pursue justice (everyone is subject to the same policies and processes) and/or decrease the political or subjective nature of evaluating individuals. Talent management asserts that employees are not of equal value. Workforce differentiation is a crucial component of talent management. The goal is to understand which specific individuals and/or roles are of greater importance (aka of higher value). Talented individuals warrant additional investment and are privy to different resources than their non-talented peers (see McDonnell and Wiblen, 2020: Chapter 5 to learn more about talent development). Talent management sees specifically designated higher-value individuals – aka talent – afforded additional resources (e.g. development opportunities, secondments) than their (perceived) lower-value workforce counterparts. Notably, higher-value individuals gain the time and attention of their respective managers and potentially direct access to the senior leadership team. Decision-makers benefit from knowing what they mean when talking about talent management or HRM, and acknowledging that the two concepts are not the same – because human resource management is about doing the same thing to everyone; talent management is about doing something (a practice, resource allocation) to someone (a specific individual).

TALENT MANAGEMENT IS ABOUT JUDGING VALUE Many individuals similarly talk about talent management as a set of practices. Talent management includes acquiring, developing and retaining talent with external stakeholders (and scholars) encouraging leaders and HR professionals to focus on the associated practices: talent acquisition, development and retention. I suggest that it is better to consider talent management as a set of value-based activities. Framing talent management around judging value is helpful because talent management is about investing in individuals who are thought (or known) to be of greater importance – aka value – for strategy execution. I define talent management as: A judgment-orientated activity, where humans make judgments about [the value of] other humans. While mediated by various contextual factors and variables (such as technology), these judgments should be informed by and aligned to current and future strategic ambitions and goals (Wiblen, 2019: 154).

A judgment-based definition recognises that actors within organisations use talent management – whether talent identification, talent development or talent retention – as a mechanism to decide which individuals warrant investment. Talent management, in this vein, includes three main attributes: (1) judgments about value;

172  Research handbook on human resource management and disruptive technologies

(2) decisions; and (3) resources. Decision-makers make judgments about the value of individuals within their workforces; relevant stakeholders then make decisions based on judgments of value; decisions about resource allocations are based on prior judgments of value. Understanding that talent management is about judging value encourages decision-makers to consider the role of human judgments and reflect on whose judgments talent management decisions will be based. Whose judgments matter the most: senior executives, line managers, HR managers, vendors, consultants?

DIGITALISATION FRAMES TALENT AS A UNIVERSAL AND STATIC CONCEPT Understanding what talent means is essential. Enacting responsible talent management strategies involves knowing the specific whos (individuals) and the specific whats (the skills and capabilities and/or pivotal roles and positions) of talent. A key question to answer is: What are the key defining characteristics of talent? Digitalisation is a mechanism to capture talent meanings. Human capital management modules (think SAP SuccessFactors, Oracle and the like) are core to enacting systematic evaluation processes. In an International Journal of Human Resource Management paper, Professor Janet H. Marler and I articulate the connection between digitalisation and talent meanings. We illuminate that systematic talent management approaches include well-defined talent definitions, evaluative criteria and ranking algorithms for generating talent lists encoded into the software. Digitalised frameworks act as a control mechanism because they submit the workforce to the same predefined what and how. Digitalisation, therefore, frames talent as a universal and static concept. Digitalisation advances universal notions of talent – a one-size-for-all understanding of talent. Digitalisation also normalises talent as a static concept because the criteria for talent are predetermined. Software code establishes boundaries around what talent is and what talent is not. These boundaries are stable. Identifying value outside of the set criteria may prove challenging. Put another way – technology takes a picture – a snapshot at one point in time in line with a one-size-fits-all understanding of the what (the characteristics) and the how (the process) of talent meanings and subsequent identification. Think of the profound difference between just taking a person’s photo and taking a video, including sound, interactions and additional contextual information. But talent meanings can and do change. Tansley’s (2011) etymology illustrates how talent meanings vary with time. It wasn’t until the nineteenth century that we started talking about talent within the context of an individual. The term is currently used as a generic term to describe an individual’s ability, accomplishments, aptitudes, brilliance, capacity, expertise, facility, flair, genius, gift, ingenuity, knack, prowess,

What decision-makers need to know about digitalised talent management  173

skill and/or strength. Talent is a concept; it is how we organise the world and the meaning we each derive from experiences. How you and I and broader society organise ourselves can and does change over time. (Love is another popular concept.) Decision-makers benefit from recognising the limitations of the static aspects arising from digitalised talent management because talent is dynamic. Framing talent as a fluid concept may minimise the negative consequences of adopting a one-way-for-all approach to defining and evaluating value.

ALGORITHMS ARE NOT OBJECTIVE Digitalised talent management involves algorithms. Algorithms provide a step-by-step procedure for accomplishing a set problem or task and assist with automating definable and sequential tasks. Algorithms, however, do not facilitate objective decision-making. We need to debunk a powerful myth: that algorithms are objective. Despite ongoing efforts of technology vendors and associated advocates in heralding algorithms as the means to overcome flawed human decision-making, the stated end goal is unachievable. We’ve been sold a pipe dream. Humans design algorithms. Therefore, algorithms, akin to humans, are based on subjective perceptions. Humans – in this case, software designers/coders – determine the criteria/parameters for consideration, the steps and the desired outcomes. The relevant humans code their perceptions of the above into the technology. The perceptions of these humans are then enacted on other humans – the workforce. Rather than eliminate biases, the perceptions and biases of these humans are rolled out and amplified within different organisational contexts. The bias in algorithms can metastasise. The prejudice and subjective perceptions of a set of specific humans (the software designers and coders) can spread from one part of an organisation to another. Vendor-embedded bias can metastasise and spread from one organisation to another. They are connected by the errors that scale – it’s easy to apply the bias between individuals (evaluators), between functions (from HR to Finance) and across organisations (from Organisation A to Organisation B). Algorithmic-based decisions can even metastasise across geographical boundaries as organisations use the same working methods in different regions (from England to Australia). Algorithms play a role in standardising decision-making criteria (the what) and processes (the why). Algorithms, however, do not, and cannot, remove bias. Algorithms cannot facilitate objective decisions because algorithms represent a particular (subjective) view of talent.

TALENT DECISIONS ARE ALWAYS SUBJECTIVE Decision-makers benefit from understanding that talent decisions are subjective.

174  Research handbook on human resource management and disruptive technologies

Whether framed as talent management or HRM, talent decisions always include humans, and humans are subjective beings. Rather than shy away from talking about the inconsistency in human decision-making, my view of talent management acknowledges that talent decisions are always subjective. Talent decisions are subjective because talent management is about judging value. Talent management involves humans, usually senior or line leaders, evaluating and judging the value of a specific individual to a task, an outcome, to the team or to current or future strategic ambitions and goals. Humans play a pivotal role. Humans, myself included, are biased beings. The Cognitive Bias Codex identifies at least 180 different ways factors and variables influence our views and judgments. We humans also have different ideas of what talent looks like. My research shows that humans view talent differently even when working in the same organisation and using the same term (Wiblen and McDonnell, 2020). Decision-makers have different ideas – subjective perceptions – about the value of effort versus outcomes. Decision-makers also have different ideas about team members. Some individuals will be described as having talent, while others will not. Maybe some individuals ‘were born with it’, while others ‘don’t have what it takes’. Elements of subjectivity prevail when we talk about potential. Potential is subjective and theoretical. We cannot know whether investing in an individual with potential will pay off at a future date. Nor can we know the specific date upon which the value is realised. Decision-makers need to understand that all talent decisions are subjective. All talent decisions include an element of bias. Decision-makers benefit from recognising that bias is everywhere. The goal is not to eliminate bias. The goal, instead, is to be more aware of the bias and partake in uncomfortable conversations about the (subjective) reasons – the why, why, why – decision-makers judge some individuals to be of greater value than others.

CONCLUSIONS AND RECOMMENDATIONS FOR ACTION Increasing digitalisation will continue to impact organisations. Current and new technological innovations will influence strategies, operational processes and key decision-makers’ decisions about how to structure and manage workforces. Regardless of the type of digitalisation or the talent-based challenge of the day, I advocate for the pursuit of solutions that recognise, rather than ignore, the complexity of the situation. I want decision-makers to improve the quality of their digitalised talent management practices and make better decisions. Better means that decisions and actions are responsible. Responsible talent decisions are: ● deliberate – undertaken with thought; ● intentional – enacted on purpose; and ● informed – based on knowledge.

What decision-makers need to know about digitalised talent management  175

Being deliberate, intentional and informed involves appreciating complexity and understanding that talent decisions are, and always will be, subjective. I recommend that decision-makers of the past, present and future work together to improve talent decisions and understand digitalisation’s positive and negative impacts, regardless of the specific form.

REFERENCES Beane, M. and Orlikowski, W. J. (2015). What difference does a robot make? The material enactment of distributed coordination. Organization Science, 26(6), 1553–73. doi:10.1287/ orsc.2015.1004. Davenport, T. H., Harris, J. G., and Shapiro, J. (2010). Competing on talent analytics. Harvard Business Review, 88(10), 53–8. Davenport, T. H. and Redman, T. C. (2020). Digital transformation comes down to talent in 4 key areas. Retrieved from https://​hbr​.org/​2020/​05/​digital​-transformation​-comes​-down​-to​ -talent​-in​-4​-key​-areas. Dery, K., Hall, R., Wailes, N., and Wiblen, S. (2013). Lost in translation? An actor-network approach to HRIS implementation. The Journal of Strategic Information Systems, 22(3), 225–37. Grant, D., Dery, K., Hall, R., Wailes, N., and Wiblen, S. (2009). Human resource information systems (HRIS): an unrealised potential. Paper presented at the Annual CIPD Centres’ Conference Nottingham, UK. Jesuthasan, R. and Boudreau, J. (2018). Reinventing Jobs: A 4-step Approach for Applying Automation to Work. Boston, Massachusetts: Harvard Business Review Press. McDonnell, A. and Wiblen, S. (2020). Talent Management: A Research Overview. Routledge. Orlikowski, W. J. and Scott, S. V. (2008). Sociomateriality: challenging the separation of technology, work and organization. The Academy of Management Annals, 2, 433 – 474. PwC (2017). 20th CEO survey: 20 years inside the mind of the CEO... what’s next? Retrieved from https://​www​.pwc​.com/​gx/​en/​ceo​-survey/​2017/​pwc​-ceo​-20th​-survey​-report​-2017​.pdf PwC (2019). Talent trends 2019: upskilling for a digital world. PwC’s 22nd Annual Global CEO Survey. Retrieved from https://​www​.pwc​.com/​gx/​en/​ceo​-survey/​2019/​Theme​-assets/​ reports/​talent​-trends​-report​.pdf PwC (2020). 2020 HR technology survey: key findings. Retrieved from https://​www​.pwc​ .com/​us/​en/​library/​workforce​-of​-the​-future/​hr​-tech​-survey​.htmlhttps://​www​.pwc​.com/​us/​ en/​library/​workforce​-of​-the​-future/​hr​-tech​-survey​.html. Schatzki, T. (2010). Materiality and social life. Nature and Culture, 5(2), 123–49. Tansley, C. 2011. What do we mean by the term ‘talent’ in talent management? Industrial and Commercial Training, 43(5), 266–74. Wiblen, S. (2016). Framing the usefulness of eHRM in talent management: a case study of talent identification in a professional services firm. Canadian Journal of Administrative Sciences, 33(2), 95–107. doi:10.1002/cjas.1378. Wiblen, S. (2019). e-talent in talent management. In M. Thite (Ed.), e-HRM: Digital Approaches, Directions and Applications (pp. 153–71). Milton Park: Routledge. Wiblen, S. and Marler, J. H. (2021). Digitalised talent management and automated talent decisions: the implications for HR professionals. The International Journal of Human Resource Management, 1–30. doi:10.1080/09585192.2021.1886149. Wiblen, S. and McDonnell, A. (2020). Connecting ‘talent’ meanings and multi-level context: a discursive approach. The International Journal of Human Resource Management, 31(4), 474–510. doi:10.1080/09585192.2019.1629988.

176  Research handbook on human resource management and disruptive technologies

Wiblen, S., Dery, K., and Grant, D. (2012). Do you see what I see? The role of technology in talent identification. Asia Pacific Journal of Human Resources, 50(4), 421–38. Wiblen, S., Grant, D., and Dery, K. (2010). Transitioning to a New HRIS: the reshaping of human resources and information technology talent. Journal of Electronic Commerce Research, 11(4), 251–67.

13. The role of disruptive technologies in talent management in Nordic multinational enterprises Violetta Khoreva, Vlad Vaiman, Tanya Bondarouk and Sari Salojärvi

INTRODUCTION Over the past two decades, talent management (TM) scholars have built an excellent body of knowledge on the relationship between TM and disruptively evolving technologies (Caligiuri et al., 2020; Vaiman et al., 2021; Wiblen, 2021). They have carefully observed the new technological developments in different TM domains, conducted research using the best available methodological approaches and offered practically relevant insights. Studies have covered the relationship between TM and disruptive technologies with a focus on specific technological artifacts such as digitalized TM (Wiblen, 2021; Wiblen and Marler, 2021), artificial intelligence (AI) (Charlwood, 2021), TM analytics (Ajunwa et al., 2017; Lazer and Radford, 2017; Minbaeva and Vardi, 2019; Zuboff, 2019; Marler and Martin, 2021) and others. Thanks to these and other scholarly efforts, substantial progress has been made in understanding how particular disruptive technologies have changed TM and its function. However, disruptive technologies evolve continuously. New technologies keep evolving, and researchers continuously set out to critically assess their influence on TM. Future scholars and practitioners will thus continue to cope with the different technological artifacts that are currently emerging (Kim et al., 2021). In this chapter, we aim to shift scholarly discovery towards an empirical exploration of the more all-encompassing expansions of disruptive technologies, and the longer-term, bigger-picture implications of technological advancement for TM and the TM function. In this context, by disruptive technologies we denote information technologies and see them as the application of digital devices to collect, store, retrieve, share, publish and disseminate data for TM processes in organizations. We trace the role of disruptive technologies in TM in general rather than focusing on a single technological artifact, thereby generating insights for future research. In addition, we accentuate the formative role of disruptive technologies in TM by illustrating the results of our qualitative study of leadership professionals working in Nordic multinational enterprises (MNEs). We advance the academic understanding of the role of disruptive technologies in TM and specify how technological advancement shapes MNEs’ TM systems. We also develop an organizational typology, which offers theoretical direction for TM scholars and practical guidance for busi177

178  Research handbook on human resource management and disruptive technologies

ness and HR leaders on how best to utilize and adopt technology to serve the goals of TM. We conclude our chapter by providing valuable insights for TM scholars and practitioners on how to navigate their new realities.

UNDERSTANDING INTERFACES BETWEEN TALENT MANAGEMENT AND TECHNOLOGY Debates on Talent Management The past decade has seen a considerable increase in focus on TM (Farndale et al., 2021; Garavan et al., 2021; Snyder et al., 2021; Vaiman et al., 2021). The TM phenomenon also remains a key area of interest for practitioners, as many continue to struggle to deliver on the talent agenda in their own company (Charan et al., 2018). While the academic scholarship has not reached a consensus on a precise definition of TM, Collings and Mellahi’s (2009) definition has been determined as the most broadly adopted (Gallardo-Gallardo et al., 2015). In particular, Collings and Mellahi (2009, p. 305) define TM as: activities and processes that involve the systematic identification of key positions which differentially contribute to the organization’s sustainable competitive advantage, the development of a talent pool of high-potential and high-performing incumbents to fill these roles, and the development of a differentiated human resource architecture to facilitate filling these positions with competent incumbents and to ensure their continued commitment to the organization.

In exploring the broader literature on TM, the literature recognizes two coexisting approaches to TM, which refer to the prevalence of talent in the organizational population (Gallardo-Gallardo et al., 2013). These approaches include the exclusive and the inclusive approaches to TM. Deeply rooted in the resource-based view of the firm (Barney, 1991) and the architectural theory of HRM (Lepak and Snell, 2002), the ‘exclusive approach’ to TM focuses on staff who contribute a disproportionate amount of output, and/or create disproportionate value, for an organization’s strategic success (Meyers and Van Woerkom, 2014; Gallardo-Gallardo, 2019). The approach is based on workforce differentiation, building on the premise that some employees are more valuable than others, and hence merit preferential treatment. In line with the exclusive approach to TM, only high performers and high potentials are identified as talents: individuals who can make a difference to organizational performance, either through immediate contributions or, in the longer-term, by demonstrating the highest levels of potential (De Boeck et al., 2018). This elite group is often known as the MNE’s talent pool. The implied exclusiveness leads to ‘rank and yank’ practices (Gallardo-Gallardo, 2019), whereby the allocation of more resources to better performers and higher potentials in the organization generates a higher return on investment, since more is invested where greater returns can be expected (Bothner et al., 2011). Thus, exclusive

The role of disruptive technologies in talent management in Nordic multinational enterprises  179

TM practices focus on attracting, recruiting, identifying, allocating, developing and retaining those individuals who demonstrate high performance and high potential, and deploying them to key positions identified beforehand. Recently, there has been a shift towards the more inclusive approach to TM, possibly due to workplace regulations requiring equal treatment of employees (Cappelli and Keller, 2014), and a scarcity of talent in the labor market (Fernández-Aráoz, 2014). Rooted in positive psychology, the ‘inclusive approach’ to TM is based on the belief that all employees have valuable strengths or talents, which, if correctly applied, can add value to the organization (Cappelli and Keller, 2014; Meyers, 2016). Hence, organizational success stems from capturing the value of the entire workforce, not just a few high performers and high potentials (Gallardo-Gallardo, 2019). Indeed, in knowledge-based organizations, all employees are considered a crucial asset to generate profits and succeed (Cascio and Boudreau, 2016). In addition, since the current labor market is highly dynamic and constantly changing, predictions about specific talents required in the future could be a ‘long guess’ (Gallardo-Gallardo, 2019). Therefore, considering different forms of talent across the workforce is seen as a more sensible strategy than just focusing on a few selected individuals (Meyers, 2016). The inclusive approach to TM is believed to benefit by treating everyone in the organization equally and, consequently, by creating a more pleasant, collegial and motivating work climate (Bothner et al., 2011). The egalitarian distribution of resources across all employees in an organization prevents loyal personnel from becoming jealous and unsatisfied if they are not provided with the same resources as high performers and high potentials (Khoreva et al., 2019). By providing everyone with the opportunity to fully unlock their talent via participation, the inclusive approach to TM is believed to enhance talent wellbeing, learning and performance (De Boeck et al., 2018). In addition, organizations that demonstrate concern for the whole personnel may find it easier to attract talent, since people prefer to work for firms that genuinely care about them (Gallardo-Gallardo, 2019). Hence, identifying an individual’s talents, encouraging their use and refinement and matching those talents with appropriate positions have become key TM tasks within this approach (Meyers and Van Woerkom, 2014; Gallardo-Gallardo, 2019). The main criticism of the inclusive approach to TM is that it makes the differentiation between TM and HRM more difficult. Referring to TM as a traditional HR function executed quickly not only makes its use ‘superfluous’ (Lewis and Heckman, 2006) but also ignores the less egalitarian and more elitist nature of TM (Collings and Mellahi, 2009). Despite some attempts to provide a unique and differential definition of the inclusive approach, the reality is that its advantages are not so clear (Gallardo-Gallardo, 2019). It is important to note, however, that one does not need to look at the exclusive-inclusive dichotomy as a divide but rather as a continuum, where both extreme approaches do exist, but where most organizations’ TM approaches nowadays fall somewhere in-between the two extremes.

180  Research handbook on human resource management and disruptive technologies

The Theorization of Disruptive Technologies There is little doubt that continuing advances in disruptive technologies will play a profound role in how TM is executed. At the same time, MNEs may react differently to the technological advancement within their TM system, as they differ in envisioning, understanding and the execution of a balance between technology and talents in generating value (Susskind and Susskind, 2015). Before entering the debate on the formative role of disruptive technologies in TM, it is crucial to define what we mean by ‘disruptive technology’ and acquaint ourselves with technology theorizations within the HRM and TM academic discourses. In this study, we define disruptive technologies as information technologies that build on and include different digital tools to collect, store, retrieve, share, publish and disseminate data for TM purposes. We emphasize that disruptive technologies cover diverse digital emerging artifacts ranging from social network sites and digital platforms to AI-based tools (Kim et al., 2021; Raisch and Krakowski, 2021). We explore the formative role of disruptive technology (Kim et al., 2021) in TM at the organizational level, where technological advancement shapes existing TM practices in new ways, offers new TM practices and replaces obsolete practices (Parviainen et al., 2017). There is a long tradition within the HRM discipline of studying the interfaces between technologies and HRM. Several reviews have analyzed technologically enabled HRM and focused on the relationships between certain technological artifacts and HRM (e.g. Bondarouk and Brewster, 2016; Bondarouk et al., 2017; Fleming, 2019; Garcia-Arroyo and Osca, 2019). In addition, Vrontis et al. (2021) have conducted a systematic review of the latest HRM developments in line with diverse technological advances. A common agreement of this scholarship is the realization that disruptive technologies can be examined from many different conceptual angles (Woodward, 1965; Leonardi and Barley, 2010). We recognize the divergence in the epistemology of disruptive technologies. Indeed, HR scholars have adopted varied approaches in conceptualizing the phenomenon of technology. In terms of diversity of technology epistemologies, we draw from the three-fold conceptualization of technology proposed by Orlikowski and Iacono (2001) that has proved to be relevant for HRM (Charlier et al., 2016). While the seminal framework for technology theorization was originally developed in the context of another discipline (information systems research), its three categories – tool, proxy and ensemble – are also highly applicable to HRM and TM research (Kim et al., 2021). Likewise, the framework itself is principally consistent with the frameworks employed in other HRM review articles (e.g. Marler and Fisher, 2013). The three theorizations of technology are described in detail below. First, the concept of disruptive technologies can be theorized in line with the ‘tool’ view of technology, which relates to the traditional understanding of technology, where it equates with a stable and determined set of equipment, procedures and techniques, purposefully designed to serve the goals of its owner (Orlikowski and Iacono, 2001). Its features, functionalities and anticipated outputs are assumed to be

The role of disruptive technologies in talent management in Nordic multinational enterprises  181

well-defined. Once adopted, the technology is intended to be under the control of the owner and thus to produce anticipated results. A complete transfer of technology between individuals or organizations is deemed plausible (Kim et al., 2021). Within this theorization of technology, HRM scholars have predicted it to be a driver of routine displacement, a means to enhance organizational productivity, a way to facilitate information processing and a tool to facilitate social relations (Kim et al., 2021). Next, the concept of disruptive technologies can be theorized as a socially constructed reality following the ‘proxy’ view of technology, which recognizes the importance of users in the adoption and implementation of technology (Orlikowski and Iacono, 2001). According to this view, technology is largely a stable entity. However, the influence of technology is understood to depend on users’ cognitive and behavioral responses. This view is different from the tool view as it acknowledges the human agency of technology adopters, who are assumed to have the ability to accept or resist technological innovations (Kim et al., 2021). Therefore, the adoption of technology does not guarantee its full acceptance or total utilization by the users (Kim et al., 2021). In the proxy view, employees’ skills and knowledge (i.e. human capital) concerning technology are assumed to be a surrogate indicator of new technology, so its adoption and implementation depends on who embodies new technology. Finally, the concept of disruptive technologies can be theorized as a part of a complex system that is shaped by various facets of organizational, social and legal environments, in line with the ‘ensemble’ view of technology. This view acknowledges the importance of the social contexts within which technology is formulated, enacted, interpreted and appropriated. It highlights the multiway interactions among technology, technology users and their surrounding organizational and institutional contexts, where technology is inseparably linked to the human agency of developers and users. Following the ensemble view of technology, scholars contextualize it into the broader technological environments across organizational, industrial and national boundaries (e.g. Francis et al., 2014). Since the ensemble view explicitly takes into account the social context around the use of technology, and operates at the societal level of technological advancement, we leave this theorization to future research on macro TM, which operates at the societal level (Vaiman et al., 2019a,b). In this chapter, we focus on the organizational level and thus spotlight only the tool and proxy views of technology.

METHODS To illustrate a comprehensive picture of the interplay between TM and technology, in this chapter, we share the results of our qualitative study among leadership professionals of 36 Nordic MNEs with diverse characteristics in terms of sector, size and number of countries of operation. In total, we conducted 36 interviews with leadership professionals with many years of experience in HR/TM, each of whom was responsible for managing talent in their MNE (see all the details in Table 13.1).

182  Research handbook on human resource management and disruptive technologies

The interviews lasted from 50 to 70 minutes and were all conducted in person in 32 headquarters of Finnish MNEs and four subsidiaries of other Nordic MNEs (two Swedish, one Norwegian and one Danish).

FINDINGS In this section, we first look at the approaches to TM that the sample MNEs follow. We then present the sample MNEs’ views of disruptive technologies. Finally, we formulate and describe in detail an organizational typology, which distinguishes MNEs based on the following criteria: TM approach, view of technology, implementation of disruptive technologies in TM practices and outcomes of disruptive technology involvement in TM. Approach to TM We found that 21 of the sample MNEs followed the more exclusive approach to TM, with clear TM systems and well-defined talent competencies. These MNEs identified only a small percentage of their employees as talents. The other 15 sample MNEs also followed the more exclusive approach to TM. However, within these MNEs, everyone had the chance to become a talent by self-initiation. The companies moved towards a ‘more flexible understanding of talent where employees can nominate themselves for the status of a talent’ (Construction 1). They favored a more open and flexible approach to TM, where employees could make a difference and apply for talent status, although the final decision was made by the manager. In other words, while the ‘traditional recipes don’t work anymore’ (Pharmaceuticals) and self-initiation was promoted, supervisors still played a decisive role in internal talent selection. Our data echo previous research on the exclusive approach to TM, discussed earlier in this chapter. Some of the sample MNEs advocated the ‘traditional’ exclusive approach to TM and viewed only a certain circle of employees as talent (Meyers and Van Woerkom, 2014; De Boeck et al., 2018; Gallardo-Gallardo, 2019). These organizations had a clear definition of what criteria a talent has to meet, in the form of attributes such as willingness to perform, ability to learn and adjust and proactivity. Other sample MNEs also followed the exclusive approach to TM, with a focus on people replete with excellent performance and potential, but with some elements of inclusive TM, which provided everyone with a chance to become a talent through self-initiation. Following this approach to TM, all employees can apply for TM programs, although talent selection is still manager-driven and follows a structured process. Harsch and Festing (2020, p. 12) describe this approach as ‘agile’, characterized by self-initiation, but nevertheless strategically reconfiguring the resource talent.

The role of disruptive technologies in talent management in Nordic multinational enterprises  183

Table 13.1 Sector

Overview of sample MNEs and respondents Origin

Position/Title

Number of

Number of

employees

countries of

Gender

Organizational type

operation Business Services Finnish

2 100

(2)

6 Senior Vice President,

F

Exclusive Tech

F

Exclusive Tech

F

Exclusive Tech

M

Exclusive Tech

16 Leadership Development M

Exclusive Tech

HR

Construction (2)

Finnish

2 583

Construction (3)

Swedish

33 585

Extractive (2)

Finnish

8 191

Financial Services Danish

20 683

45 Senior Vice President,

Master

HR 17 Executive Vice

Master

President, HR 15 Senior Vice President,

Master

HR (2)

Master

Director

Financial Services Finnish

249

(3)

4 Head of HR and Legal

Master F

Exclusive Tech

F

Exclusive Tech

Affairs

IT, Electronics & Finnish

5 097

4 Executive Vice

Master

President, HR

Telecomm-

Master

unications (2) Manufacturing

Finnish

55 075

60 Head of Competence

F

Management

and Engineering

Exclusive Tech Master

(1) Manufacturing

Finnish

18 000

70 Executive Vice

F

President, HR

and Engineering

Exclusive Tech Master

(4) Manufacturing

Finnish

11 552

4 Vice President, HR

F

Exclusive Tech Master

and Engineering (6) Manufacturing

Finnish

15 000

100 Vice President, Talent

F

and Management

and Engineering

Exclusive Tech Master

(8) Manufacturing & Finnish

3 380

Engineering (10)

4 Senior Vice President,

F

Exclusive Tech

HR and Communication

Manufacturing & Finnish

Master

4 881

45 Vice President, HR

F

Exclusive Tech

2 473

11 Senior Vice President,

F

Exclusive Tech

Engineering (11) Retail and

Master Finnish

HR Development

Consumer Goods

Master

(6) Retail and

Finnish

24 000

5 TM Lead

F

Exclusive Tech Master

Consumer Goods (7) Retail and

Finnish

7 304

30 Director, People,

F

Processes and Culture

Consumer Goods

Exclusive Tech Master

(9) Business Services Finnish

2 250

(1) Construction (1)

4 People and Culture

F

Flexible Tech

F

Flexible Tech

Manager Swedish

16 000

11 Senior Vice President, HR and Communications

Master Master

184  Research handbook on human resource management and disruptive technologies Sector

Origin

Position/Title

Number of

Number of

employees

countries of

Gender

Organizational type

operation Financial Services Finnish

12 269

(1)

16 Leadership Development M Manager

Infrastructure and Finnish

Master

7 500

4 Talent Lead

F

Flexible Tech

24 455

35 Talent Lead

F

Flexible Tech

26 Head of Talent

F

Flexible Tech

Utilities

Master

Manufacturing & Finnish Engineering (2) Manufacturing

Flexible Tech

Master Finnish

4 018

Management

and Engineering

Master

(9) Retail and

Norwegian

12 883

50 Vice President, HR

F

Flexible Tech Master

Consumer Goods (3) Retail and

Finnish

9 734

5 Talent Lead

M

Flexible Tech Master

Consumer Goods (5) Retail and

Finnish

4 790

60 Head of People and

F

Culture

Consumer Goods

Flexible Tech Master

(8) Transport

Finnish

6 788

127 Senior Vice President,

F

Flexible Tech

F

Exclusive Tech

HR IT, Electronics & Finnish

24 000

20 Executive Vice

Master

President, HR

Telecomm-

Companion

unications (1) IT, Electronics & Finnish

750

20 Senior Vice President,

F

HR

Telecomm-

Exclusive Tech Companion

unications (3) Manufacturing & Finnish

18 700

12 Vice President, HR

F

Exclusive Tech

13 600

30 Senior Vice President,

F

Exclusive Tech

3 491

24 Senior Vice President,

F

Exclusive Tech

Engineering (3)

Companion

Manufacturing & Finnish Engineering (7) Retail and

HR Finnish

Companion

HR

Consumer Goods

Companion

(4) Extractive (1)

Finnish

4 400

14 Director, HR

F

Flexible Tech

12 Vice President, HR

F

Flexible Tech

30 Senior Vice President,

F

Flexible Tech

F

Flexible Tech

Development Pharmaceuticals

Finnish

3 446

Manufacturing & Finnish

17 076

Companion Companion

Engineering (5) Retail and

HR Finnish

22 476

8 Director, Talent

Consumer Goods

and Organizational

(1)

Development

Retail and

Finnish

13 242

9 Head of People

Consumer Goods

Development and

(2)

Wellbeing

Companion Companion F

Flexible Tech Companion

The role of disruptive technologies in talent management in Nordic multinational enterprises  185

Disruptive Technologies and TM We discovered that 26 of the sample MNEs followed the tool view of technology. They saw technology as a necessary, functional tool for maintaining various practices, including TM. The respondents in these MNEs described technology as an assistant or toolkit for mastering TM practices, including optimizing, systematizing and improving talent attraction, development and retention practices. In short, these MNEs added mainly technological enhancements to their existing TM practices. The outcomes of technology involvement in TM were the following. First, these MNEs perceived technology as an efficiency driver of the TM function. As such, they implemented technological artifacts that increased the speed, automatization and systematization of TM practices. Second, these MNEs depicted disruptive technologies as a means to decrease the influence of distance on the TM function. They implemented technological artifacts for virtual collaboration and encouraged talents to become frequent users of the artifacts in order to build a ‘we together’ (Retail and Consumer Goods 6) culture. Finally, these MNEs portrayed disruptive technologies as a way to facilitate the execution of better and more accurate TM decisions. As such, the companies installed TM analytics programs aimed at executing smarter and faster TM decisions. In contrast, the other 10 MNEs in our sample followed the proxy view of technology. They did not treat technology as a tool to support the TM function. Instead, according to the respondents, the companies perceived technology as a ‘companion aiming to transform the mindset of talents from new is frightening to new is a treasure’ (Retail and Consumer Goods 2). The respondents described disruptive technologies as an agent, coach, expert, partner, digital gentleman, or teammate, who aimed to bring out the human factor among talents, to ‘inspire them to experiment with virtual reality and find “wow” experiences’ (Retail and Consumer Goods 2). The outcomes of technology involvement in TM were the following. First, these MNEs regarded technology as a means to improve the quality of the talents’ work, encouraging them to make full use of their knowledge, skills and abilities when working with technological artifacts. Furthermore, talents were encouraged to use science, data and technological artifacts in tandem with their own capabilities, which would enable them to come up with better solutions. Second, these MNEs portrayed technology as a way to facilitate innovative and learning behaviors among talents. The companies recognized that technology needed to operate side by side with talents, and identified it as a supportive companion, a new source of value. The implementation of technological artifacts in these MNEs facilitated talents’ personal growth, self-management, self-renewal and self-exploration. Our empirical study confirmed the previous research on the views of technology, discussed earlier in this chapter. In particular, some MNEs hold the tool view of technology, where disruptive technologies are intended to be under the control of the owner and thus to produce anticipated results (Orlikowski and Iacono, 2001). In contrast, other MNEs follow the proxy view of technology, which recognizes the importance of users in technology adoption and implementation (Orlikowski and

186  Research handbook on human resource management and disruptive technologies

Iacono, 2001) and acknowledges the human agency of technology adopters, who are assumed to have the ability to accept or resist technological innovations (Kim et al., 2021). In accordance with the proxy theorization, these MNEs focus on disruptive technologies as viewed by the talents themselves, in line with talents’ perceptual, cognitive and attitudinal responses towards technology. These MNEs conceptualize disruptive technologies as a supportive companion, with whom talents experiment and realize their full potential. The companies see disruptive technologies as a means to improve the talents’ quality of work, and a way to facilitate innovative and learning behaviors among talents. Organizational Typology Based on our analysis, we formulate an organizational typology (see Figure 13.1), which distinguishes organizations based on the following criteria: approach to TM, view of technology, implementation of technology in TM practices and outcomes of technology involvement in TM. We thus identified four distinct organizational types: ‘Exclusive Tech Master’, ‘Flexible Tech Master’, ‘Exclusive Tech Companion’ and ‘Flexible Tech Companion’. We unpack each quadrant of the typology in the following subsections.

Figure 13.1

Organizational typology

Exclusive Tech Master Exclusive Tech Masters tend to develop talent pools of high-potential and high-performing employees, who fill key positions that differentially contribute to organizational competitive advantage. Hence, these organizations’ typology carries the word ‘exclusive’ in its title, even though the actual techniques may still include some inclusive components (see our earlier note on the continuum nature of the exclusive/inclusive ‘divide’). Exclusive Tech Masters view disruptive technologies

The role of disruptive technologies in talent management in Nordic multinational enterprises  187

as a tool that ‘masters’ the enhanced operational excellence of the TM function, again expressed in the typology. In particular, the involvement of disruptive technologies in TM is intended to reduce costs associated with the TM function, increase the efficiency of the function, provide faster communication and virtual collaboration and establish more accurate TM decision-making. For this reason, technological artifacts are used to optimize and systematize talent attraction, selection, identification and evaluation practices. TM analytics are also implemented. Finally, digital tools for talent raking and performance programs are employed. Flexible Tech Master Flexible Tech Masters also follow the more exclusive approach to TM, with a focus on people who demonstrate excellent performance and potential. Further, Flexible Tech Masters encourage self-initiation amongst their employees, enabling them to apply for TM programs on their own behalf. While the decision on talent status is made by a manager/supervisor, it leaves room for more TM inclusivity through flexibility and self-initiation. Hence, these organizations’ typology carries the word ‘flexible’ in its title. As is the case with Exclusive Tech Masters, Flexible Tech Masters view disruptive technologies as a tool, one which masters, that is, to say leads, to intended outcomes under the control of its owners. The word ‘master’ appears in the title. However, in contrast to Exclusive Tech Masters, Flexible Tech Masters perceive disruptive technologies as a necessary means to compete with other organizations in finding and securing the best available talents. These organizations see the necessity of implementing advanced technologies to compete with others in attracting, developing and retaining the best people, by providing them with the latest and most advanced technological solutions. In doing so, Flexible Tech Masters practice recruitment via social media, emphasize digital employer branding and encourage the use of smart devices for training, mentoring and coaching. These organizations also actively adopt digital wellbeing tools. Exclusive Tech Companion As is the case with Exclusive Tech Masters, Exclusive Tech Companions develop talent pools of high-potential and high-performing employees, who fill key positions that differentially contribute to organizational competitive advantage. Hence, these organizations’ typology carries the word ‘exclusive’ in its title. However, Exclusive Tech Companions amplify the proxy view of technology, and exert agency over how disruptive technologies are deployed by talents. These organizations express a proactive stance on technology, conceptualizing it as something that talents and organizations do through everyday practice and social interaction. Exclusive Tech Companions treat technology as a supportive companion, with whom talents experiment and realize their full potential. The organizations encourage talents to make full use of their knowledge, skills and abilities when partnering with disruptive

188  Research handbook on human resource management and disruptive technologies

technologies, in order to improve the quality of the talents’ work and enable creative, nonstandardized behaviors amongst them. Flexible Tech Companion In the same way as Flexible Tech Masters, Flexible Tech Companions follow the more exclusive approach to TM, with a focus on people delivering excellent performance and potential. The organizations encourage self-initiation amongst their employees, who can themselves apply for TM programs. While this approach is exclusive – that is, the decision on talent status is made by a manager/supervisor – it does leave some room for flexibility and self-initiation. Flexible Tech Companions exercise the proxy view of technology and exert agency over how disruptive technologies are deployed by talents. The organizations treat technology as a supportive companion, with which talents experiment and realize their full potential, and ‘companion’ appears in the title. Flexible Tech Companions create unique ecosystems where talents and technology are deeply intertwined, and thus, generate unique value. In these organizations, disruptive technologies inspire talents, offering a platform for personal growth, self-management, self-renewal and self-exploration. Flexible Tech Companions seek like-minded talents who are willing to work closely with disruptive technologies. They focus on deep-level discussions with talents attempting to fulfill their individual potential and needs.

CONCLUSIONS AND RECOMMENDATIONS FOR ACTION This chapter has aimed at advancing the latest discourse on technology, by showing a discursive diversity of understanding of disruptive technologies for TM purposes that organizations employ. Specifically, we illustrated that many organizations hold the tool view of technology, conceptualizing disruptive technologies as a productivity-enhancing mechanism, which leads to the increased efficiency of the TM function, decreased influence of distance on the function and better TM decision execution. Hence, in these organizations, any new disruptive technology ‘at best, improve[s] how companies do what they’ve always done’ (Ross et al., 2019, p. 144). Other organizations follow the proxy view of technology. They perceive disruptive technologies as an intelligent supportive companion with a new source of value, which possibly creates new opportunities for talents and organizations in general. Further, they focus on disruptive technologies as viewed by the talents themselves, in line with talents’ perceptual, cognitive and attitudinal responses to the technology. These organizations see disruptive technologies as a means to improve talents’ work quality, facilitate innovative and learning behaviors amongst them and for unique value creation and talent growth (Brynjolfsson and McAfee, 2017). By showing the diversity of view on technology that organizations hold, we also accentuate that organizations apply technologies differently for TM purposes.

The role of disruptive technologies in talent management in Nordic multinational enterprises  189

This chapter has also offered an organizational typology, which represents a theoretical direction to explore a formative role for disruptive technologies in TM. The typology distinguishes four different organizational types based on the following criteria: approach to TM, view of technology, implementation of technology in TM practices and outcomes of technology involvement in TM. For Exclusive Tech Masters, the formative role of disruptive technologies in TM lies in the enhanced operational excellence of the TM function. Flexible Tech Masters see the seminal role of disruptive technologies in TM as attracting, developing and retaining the best people by providing them with the latest, most advanced technological solutions. For Exclusive Tech Companions, the overall role of disruptive technologies in TM includes supporting talents to experiment and realize their full potential by working in tandem with technology. Finally, for Flexible Tech Companions, the formative role of disruptive technologies in TM incorporates encouraging talents to focus on new value-creating activities by partnering with technology. This chapter has offered some practical advice for business leaders and HR professionals. We have formulated an organizational typology that can serve as a guide for those responsible for the TM function in their organization. Each organizational type is elucidated in detail in previous sections and is illustrated in Figure 1, so organizations can easily recognize their position and realize what needs to be done. Below, we provide guidance for organizations on what they need to accomplish, if they choose to pursue a particular aspect of the typology. If an organization intends to differentiate itself as an Exclusive Tech Master (i.e. pursues a more exclusive approach to TM and views disruptive technologies as a tool that enhances operational excellence of the TM function), it needs to invest heavily in the technological advancement of their TM function, including automation and standardization of attraction, selection, identification and performance practices, and the implementation of TM data analytics. If an organization intends to act as a Flexible Tech Master, recruitment via social media, digital employer branding, mentoring and coaching, which stipulate the use of the latest technological innovations, should be the essential factors in employee acquisition and retention. The organization should also focus on meeting the individual needs of talents and deriving the appropriate solutions for them, assisted by disruptive technologies. Finally, technological artifacts that aid the enhancement of talents’ wellbeing need to be employed. If an organization intends to operate as an Exclusive Tech Companion, it needs to encourage talents to work closely with technology, so that machine abilities and talents’ unique capabilities including, for instance, commonsense reasoning and intuition, evolve in tandem. Through continuous experimentation with disruptive technologies, talents and technology should intertwine to a degree where they can collectively exhibit entirely new, emergent behaviors and value-creation activities that neither can achieve individually. Should an organization intend to distinguish itself as a Flexible Tech Companion, it needs to seek out like-minded talents with excellent digital and analytical skills, who are not afraid of technology, of stepping outside their comfort zone, or of working with disruptive technologies instead of handing tasks over to technologies.

190  Research handbook on human resource management and disruptive technologies

The organization needs to invest heavily in the integration of technology into the overall organizational ecosystem. Self-management and self-exploration among talents should be prioritized. Finally, the organization needs to focus on deeper-level discussions with talents in attempting to meet their individual needs. Yet, whatever approach to TM organizations opt for, and whatever organizations decide to do regarding their use of disruptive technologies, it is important to remember that no one scenario is necessarily better or more successful than another. TM strategy in general, and the approach to TM in particular, is determined by a company’s strategic intent, its size and age, the industry involved and even the prevailing national culture of the organization’s location. This means anything organizations do in terms of TM, including its technological advancement, should first and foremost depend upon the overall organizational strategy (Vaiman et al., 2012). In other words, any organization needs first to decide on its strategy, next on its approach to TM and only then deal with the desired role of disruptive technologies in its TM function.

REFERENCES Ajunwa, I., Crawford, K. and Schultz, J. (2017) Limitless worker surveillance. California Law Review, 105(3): 735–76. Barney, J.B. (1991) Firm resources and sustained competitive advantage. Journal of Management, 17: 99–120. Bondarouk, T. and Brewster, C. (2016) Conceptualising the future of HRM and technology research. The International Journal of Human Resource Management, 27(21): 2652–71. Bondarouk, T., Parry, E. and Furtmueller, E. (2017) Electronic HRM: four decades of research on adoption and consequences. International Journal of Human Resource Management, 28(1): 98–131. Bothner, M.S., Podolny, J.M. and Smith, E.B. (2011) Organizing contest for status: the Matthew effect vs the Mark effect. Management Science, 57(3): 439–57. Brynjolfsson, E. and McAfee, A. (2017). The business of artificial intelligence. Harvard Business Review, 7(1), 1–2. Caligiuri, P., De Cieri, H., Minbaeva, D., Verbeke, A. and Zimmermann, A. (2020) International HRM insights for navigating the COVID-19 pandemic: implications for future research and practice. Journal of International Business Studies, 51: 697–713. Cappelli, P. and Keller, J. (2014) Talent management: conceptual approaches and practical challenges. Annual Review of Organizational Psychology and Organizational Behaviour, 1(1): 305–31. Cascio, W.F. and Boudreau, J.W. (2016) The search for global competence: from international HR to talent management. Journal of World Business, 51(1): 103–14. Charan, R., Barton, D. and Carey, D. (2018) Talent Wins: The New Playbook for Putting People First. Harvard Business Press. Charlier, S.D., Guay, R.P. and Zimmerman, R.D. (2016) Plugged in or disconnected? A model of the effects of technological factors on employee job embeddedness. Human Resource Management, 55(1): 109–26. Charlwood, A. (2021) Artificial intelligence and talent management. In S. Wiblen (ed.), Digitalised Talent Management: Navigating the Human-Technology Interface (1st ed.). Routledge, 122–36. Collings, D.G. and Mellahi, K. (2009) Strategic talent management: a review and research agenda. Human Resource Management Review, 19(4): 304–13.

The role of disruptive technologies in talent management in Nordic multinational enterprises  191

De Boeck, G., Meyers, M. and Dries, N. (2018) Employee reactions to talent management: assumptions versus evidence. Journal of Organizational Behavior, 39(2): 199–213. Farndale, E., Thite, M., Budhwar, P. and Kwon, B. (2021) Deglobalization and talent sourcing: cross-national evidence from high-tech firms. Human Resource Management, 60: 259–72. Fernández-Aráoz, C. (2014) 21st century talent spotting. Harvard Business Review, 92: 46–56. Fleming, P. (2019) Robots and organization studies: why robots might not want to steal your job. Organization Studies, 40(1): 23–38. Francis, H., Parkes, C. and Reddington, M. (2014) E-HR and international HRM: a critical perspective on the discursive framing of e-HR. The International Journal of Human Resource Management, 25(10): 1327–50. Gallardo-Gallardo, E. (2019) The meaning of talent in the world of work. In D.G. Collings, H. Scullion, and P.M. Caligiuri (eds), Global Talent Management (2nd ed.). Routledge, 33–58. Gallardo-Gallardo, E., Dries, N. and Gonzalez-Cruz, T.F. (2013) What is the meaning of ‘talent’ in the world of work?. Human Resource Management Review, 23: 290–300. Gallardo-Gallardo, E., Nijs, S., Dries, N. and Gallo, P. (2015) Towards an understanding of talent management as a phenomenon-driven field using bibliometric and content analysis. Human Resource Management Review, 25(3): 264–79. Garavan, T.N., Morley, M.J., Cross, C., Carbery, C. and Darcy, C. (2021) Tensions in talent: a micro practice perspective on the implementation of high potential talent development programs in multinational corporations, Human Resource Management, 60: 273–93. Garcia-Arroyo, J. and Osca, A. (2019) Big data contributions to human resource management: a systematic review. The International Journal of Human Resource Management, 32(10): 1–26. Harsch, K. and Festing, M. (2020) Dynamic talent management capabilities and organizational agility – a qualitative exploration. Human Resource Management, 59: 43–61. Khoreva, V., Vaiman, V. and Kostanek, E. (2019) Talent identification transparency: an alternative perspective. European Journal of International Management, 13(1): 25–40. Kim, S., Wang, Y. and Boon, C. (2021) Sixty years of research on technology and human resource management: looking back and looking forward, Human Resource Management, 60(1): 229–47. Lazer, D. and Radford, J. (2017) Data ex machina: introduction to big data. Annual Review of Sociology, 43: 19–39. Leonardi, P.M. and Barley, S.R. (2010) What’s under construction here? Social action, materiality, and power in constructivist studies of technology and organizing. Academy of Management Annals, 4(1): 1–51. Lepak, D.P. and Snell, A.A. (2002) Examining the human resource architecture: the relationships among human capital, employment, and human resource configurations. Journal of Management, 28(4): 517–43. Lewis, R.E. and Heckman, R.J. (2006) Talent management: a critical review. Human Resource Management Review, 16(2): 139–54. Marler, J.H. and Fisher, S.L. (2013) An evidence-based review of e-HRM and strategic human resource management. Human Resource Management Review, 23(1): 18–36. Marler, J.H. and Martin, L. (2021) People analytics maturity and talent management: linking talent management to organisational performance. In S. Wiblen (ed.), Digitalised Talent Management: Navigating the Human-Technology Interface (1st ed.). Routledge, 79–97. Meyers, M.C. and Van Woerkom, M. (2014) The influence of underlying philosophies on talent management: theory, implications for practice, and research agenda. Journal of World Business, 49(2): 192–203. Meyers, M.C. (2016) Talent management: towards a more inclusive understanding. Tijdschrift Voor HRM, 12: 1–13. Minbaeva, D. and Vardi, S. (2019) Global talent analytics. In D.G. Collings, H. Scullion, and P.M. Caligiuri (eds), Global Talent Management (2nd ed.). Routledge, 179–200.

192  Research handbook on human resource management and disruptive technologies

Orlikowski, W.J. and Iacono, C.S. (2001) Research commentary: desperately seeking the ‘IT’ in IT research – call to theorizing the IT artifact. Information Systems Research, 12(2): 121–34. Parviainen, P., Tihinen, M., Kääriäinen, J. and Teppola, S. (2017) Tackling the digitalization challenge: how to benefit from digitalization in practice. International Journal of Information Systems and Project Management, 5(1): 63–77. Raisch, S. and Krakowski, S. (2021) Artificial intelligence and management: the automation-augmentation paradox. Academy of Management Review, 46(1): 192–210. Ross, J.W., Beath, C.M. and Mocker, M. (2019) Designed for Digital. How to Architect Your Business for Sustained Success. MIT Press. Snyder, D.G., Stewart, V.R. and Shea, C.T. (2021) Hello again: managing talent with boomerang employees. Human Resource Management, 60: 295–312. Susskind, R. and Susskind, D. (2015) The Future of The Professions: How Technology Will Transform the Work of Human Experts. Oxford University Press. Vaiman, V., Collings, D. and Scullion, H. (2012) Talent management decision making. Management Decision, 50(5): 925–41. Vaiman, V., Sparrow, P., Schuler, R. and Collings, D.G. (2019a) Macro Talent Management. A Global Perspective on Managing Talent in Developed Markets. Taylor & Francis. Vaiman, V., Sparrow, P., Schuler, R. and Collings, D.G. (2019b) Macro Talent Management in Emerging and Emergent Markets. A Global Perspective. Taylor & Francis. Vaiman, V., Cascio, W.F., Collings, D.G, and Swider, B.W. (2021) The shifting boundaries of talent management. Human Resource Management, 60: 253–7. Vrontis, D., Christofi, M., Pereira, V., Tarba, S., Makrides, A. and Trichina, E. (2021) Artificial intelligence, robotics, advanced technologies and human resource management: a systematic review. The International Journal of Human Resource Management. DOI: 10.1080/09585192.2020.1871398. Wiblen, S. (2021) Digitalised talent management: an introduction. In S. Wiblen (ed.), Digitalised Talent Management: Navigating the Human-Technology Interface (1st ed.). Routledge. Wiblen, S. and Marler, J.H. (2021) Digitalised talent management and automated talent decisions: the implications for HR professionals. The International Journal of Human Resource Management, 32(12): 2592–621. Woodward, J. (1965) Industrial Organization: Theory and Practice. Oxford University Press. Zuboff, S. (2019) The Age of Surveillance Capitalism: The Fight for a Human Future at the New Frontier of Power. Public Affairs.

14. Hiring algorithms: redefining professional roles with artificial intelligence Elmira van den Broek, Anastasia Sergeeva and Marleen Huysman

INTRODUCTION Human resource management (HRM) activities such as recruitment, selection and appraisal are increasingly supported by artificial intelligence (AI). Contemporary AI tools are powered by machine learning algorithms that derive and adapt patterns from large amounts of data to create new insights and support decision-making (Faraj et al., 2018; Vrontis et al., 2022). For instance, algorithms may uncover criteria from data that are important in selecting candidates but unknown to recruiters (Van den Broek et al., 2021), or are expected to reduce ‘noise’ or variability in employment decisions by automatically executing decision rules with unprecedented consistency (Houser, 2019; Kahneman et al., 2021). With the potential to augment or even automate human resources (HR) decisions, the use of AI is expected to have profound implications for the role and position of HR professionals that are yet to be uncovered (Meijerink et al., 2021). In this chapter, we seek to unpack the practices through which HR professionals engage with AI tools and the consequences for their work roles in organizations. AI can be considered a disruptive technology for the HR profession in the sense that it may radically change the activities that HR professionals perform on a daily basis (Vrontis et al., 2022). However, the consequences of these changes for work roles are yet to be uncovered, which is what we set out to explore in this chapter. In the following sections, we first describe how the use of technology relates to the profession’s historical desire for a more prominent role in organizations, in which AI has been framed as the latest ‘opportunity’ for the profession. Drawing on evidence from an in-depth field study of an HR department at a large international company that introduced an AI hiring tool, we then demonstrate how the technology provided an occasion for HR professionals to reshape their roles in the hiring process. Rather than interpreting the tool as a threat to their expertise, HR professionals embraced it as an opportunity to elevate their position. However, in the process of engaging with the technology, their roles increasingly centred on serving the demands of the algorithm. As a result, HR professionals gained influence over senior managers, who valued the outputs produced by the AI tool, while losing the opportunity to develop and apply their unique expertise. These potential outcomes raise new concerns for the future of the HR profession. 193

194  Research handbook on human resource management and disruptive technologies

HR’S PURSUIT OF A STRATEGIC ROLE The status of HR as a managerial profession has been a recurring concern over time, in which HR professionals have been criticized for their traditional administrative and compliance role in the workplace (Beer, 1997; Ulrich, 1998; Truss et al., 2002; Wright, 2008; Heizmann and Fox, 2019; Sandholtz et al., 2019). HR professionals have been repeatedly admonished to reexplore their role, value and competencies within organizations. Some have emphasized the need for the HR profession to focus on its ethical and social contribution by balancing employee and organizational interests (Legge, 1995; Kochan, 2004). Others have proposed a strategic role for HR that demonstrates measurable contributions to organizational performance (Beer, 1997; Ulrich, 1998; Lawler and Mohrman, 2003; Ulrich and Brockbank, 2005). This strategic HRM movement has fuelled prodigious practitioner and academic streams of literature with a long tradition of research exploring the role of technology in strengthening the position of HR within organizations (Kovach et al., 2002; Marler and Parry, 2016; Bondarouk et al., 2017; Marler and Boudreau, 2017). The latest wave of HRM research has praised new technologies such as AI for broadening the strategic influence and respectability of HR (Ulrich and Dulebohn, 2015; Margherita, 2021; Budhwar et al., 2022; Vrontis et al., 2022). These technologies are viewed as an opportunity for HR professionals to play a more strategic role in organizations by making decisions based on measurable ‘facts’ rather than intuition and gut feeling. For example, Pessach and co-authors (2020) demonstrated that a hybrid decision-support tool may support and benefit the recruiters’ decision-making process by revealing counterintuitive insights and biased recruitment practices. These claims echo prior promises with regard to the strategic benefits of technology-enabled HRM. Scholars have discussed a variety of tools such as HR metrics, benchmarking, HR scorecards, evidence-based management and e-HRM for decades without a noticeable change in the HR position (Rasmussen and Ulrich, 2015). The rise of AI tools based on advanced machine learning techniques has renewed attention on the use of technology to improve HR decisions and facilitate a more strategic role for the HR profession (Vrontis et al., 2022). Others have warned that such technologies may pose unique challenges for the HR profession. For example, machine learning algorithms place great demands on high-quality data that is difficult to achieve in HR settings (Tambe et al., 2019), are opaque to workers and developers (Burrell, 2016) and raise various ethical and legal concerns (Tursunbayeva et al., 2022). Rather than strengthening the status of HR, these technologies might undermine professionals by marginalizing their judgment in decision-making (Greasley and Thomas, 2020; Meijerink et al., 2021; Wiblen and Marler, 2021). Wiblen and Marler (2021) showed, for instance, how a digital talent management technology limited HR’s involvement in talent identification by devaluing their professional judgment and dictating decisions. When HR professionals embraced the tool in their work, their prominent role in talent identification shifted towards an administrative role centred on defending the managers’ decisions and outputs of the algorithm. This work, therefore, cautions about the often-stated pos-

Hiring algorithms  195

itive relationship between the use of technology and HR’s strategic influence in the organization (Angrave et al., 2016; Marler and Boudreau, 2017; Cheng and Hackett, 2021; Meijerink et al., 2021; Wiblen and Marler, 2021). Scholars have only recently started discussing the implications of AI tools for the HR profession. These studies have shed light on the opportunities and challenges that surround AI solutions in HRM and how they relate to the distinctive characteristics of AI tools (Margherita, 2021; Meijerink et al., 2021; Budhwar et al., 2022; Vrontis et al., 2022). However, we need to know more about what happens to HR roles on the ground when AI tools enter the organization. To explore the activities that are involved in and contribute to changes in HR roles, this chapter focuses on the everyday practices through which people and technology interact (Orlikowski, 2000). Using such a situated practice lens helps to recognize that, to understand shifts in roles, we need to attend to the recurrent activities co-constituted by both professionals and technology through which roles are enacted. It is this approach that informed our study of the introduction of an AI tool in the hiring process.

METHODS The Setting of AI-Based Hiring We draw on findings from an in-depth field study at the HR department of a large international company (MultiCo1) that introduced an AI tool in the hiring process. There has been a rapidly growing interest in the use of algorithms in various stages of the hiring process, including in the sourcing, screening and interviewing of job candidates (Budhwar et al., 2022). These tools are often accompanied by claims around cost and time savings but also the promise of diminishing human bias in decision-making, thereby increasing objectivity and fairness in the hiring process (Köchling and Wehner, 2020). This study particularly focuses on the use of AI for screening and interviewing purposes with the intent to ‘empower [HR and senior manager] teams to make better selection decisions’ (field notes). At the time of data collection, MultiCo belonged to one of the world’s largest fast-moving consumer goods companies, with almost 200,000 employees in more than 50 countries. The introduction of AI resonated with the strong data-driven culture of the company, in which employees were encouraged to base decisions on data and ‘facts’ rather than personal experience or intuition. As a senior manager joked: ‘You always must back up decisions with data. Even if the data does not exist, people will still ask for it!’ (field notes). Being familiar with the ongoing discussions about the potential of AI techniques in the HR community, the HR department understood the technology as an opportunity to improve decision-making and elevate the HR function. In 2018, the HR department introduced an AI tool in the hiring process. The tool was designed by an external vendor and relied on machine learning and over 50 game- and video-based assessments to offer insight into people’s work potential.

196  Research handbook on human resource management and disruptive technologies

The game-based assessments aimed to measure personality traits (e.g. extraversion), skills (problem-solving), and career values (e.g. work-life balance) in an interactive gaming environment. In the video-based assessments, candidates were asked to record short answers to questions about themselves, such as ‘What is the most effective way for you to learn something new?’ in which data on verbal and nonverbal communication and facial expressions was collected. This data was fed to a machine learning algorithm to predict job candidates’ fit with the company based on their overlap with the attributes of existing high-performing employees. Table 14.1 summarizes the change the AI tool brought to the company’s existing way of hiring. Previously, HR professionals screened candidates’ resumes, motivation letters and IQ test scores by relying on their professional judgment; after which, senior managers interviewed candidates based on input provided by the HR team and made final hiring decisions. With the tool, the AI predictions on candidates became leading in determining screening decisions; after which, senior managers interviewed candidates with the help of AI outputs and made final hiring decisions. Table 14.1

Overview of the hiring process before and with the use of the AI tool

Stage

Before the use of the AI tool

With the use of the AI tool

Screening by HR

Using professional judgment informed by:

Using AI predictions informed by:

professionals

- Resumes

- Data from online games

- Motivation letters

- Data from video recordings

- IQ test scores Interviewing by senior

Using professional judgment informed by:

Using professional judgment informed by:

managers

- HR method and suggestions

- AI outputs

Data Collection and Analysis This chapter draws on data collected during a three-year field study on the development and use of an AI hiring tool.2 The focus of this chapter lies on the work of HR professionals involved with the tool. The data includes observations, interviews and archival data. Observations focused on the work practices of HR professionals, including activities such as screening job candidates with or without the aid of the AI tool, preparing AI outputs and training senior managers. The researcher shadowed HR professionals during the day and attended internal team meetings and sessions with other stakeholders, such as job interviews with candidates and meetings with the people analytics team. Detailed notes made during the day were expanded into written field notes. We also carried out interviews with members of the HR department, including HR professionals, senior managers and data scientists. These interviews specifically focused on the perceptions of HR’s activities in the hiring process before the introduction of the tool and HR’s involvement with AI in the hiring process as the

Hiring algorithms  197

tool was implemented. Additionally, secondary documents were collected, including company reports, PowerPoints and screenshots of AI scores and outputs. We analysed our data by following a process research approach (Langley, 1999), in which we traced the flow of events and changes in activities with the introduction of the AI tool. By carefully examining the nature of the changes in activities, the reasons for them and consequences for HR professionals, we came to conceptualize the shift in activities as changes to the roles of HR professionals in the hiring process. We found that these shifts had important implications for HR’s position in the hiring process, in which HR professionals lost decision-making authority to the algorithm but shaped hiring decisions in indirect ways through their engagement in new activities around the algorithm.

CHANGING HR ROLES WITH AI We report on the implications of the AI tool for the areas of work in which HR professionals showed the main shifts in their work roles and activities. These involved HR’s role as gatekeeper, guardian, and mediator in the hiring process, as summarized in Table 14.2. Table 14.2

Overview of HR’s roles, shifts, and activities before and with the AI tool

HR role

Activities before AI

Activities with AI

Shift in role

Gatekeeper

HR professionals decide

HR professionals supply and

From selecting candidates for

which candidates to present

select data for the algorithm so

managers to selecting data for

to managers based on their

that it can generate predictions

the algorithm.

professional judgment.

on candidates.

HR professionals guard fair

HR professionals guard fair

From guarding fair hiring

hiring by training managers

hiring by explaining and

based on expert methods to

with help of expert methods.

transforming AI outputs to

guarding fair hiring based on

managers.

AI outputs.

HR professionals mediate

HR professionals mediate

From mediating interpersonal

disputes around human

disputes around human and AI

disputes to mediating

assessment.

assessment.

human-AI disputes.

Guardian

Mediator

Gatekeeper: From Selecting Candidates to Selecting Data A first shift involved a change in HR’s gatekeeping role, in which HR professionals gave up control over deciding which candidates to present to managers and selected data for the algorithm instead. Before the use of AI, HR professionals played a key role in selecting candidates by assessing candidates’ resumes, motivation letters and IQ test scores based on their professional judgment. Candidates who were judged to be a good fit with the company were selected for job interviews with senior manag-

198  Research handbook on human resource management and disruptive technologies

ers. With the new AI tool, candidate selection was promised to become much more efficient and objective by having an algorithm infer selection criteria from data and execute these criteria without the need for human intervention. The HR team strongly believed in the potential of the AI tool to make rigorous and objective decisions and was therefore willing to rely on the tool as the sole mechanism for screening candidates. Why AI? We truly believe it’s one of the most unbiased ways of assessing candidates. So, we don't look at their resumes, prior experience, gender, or background. We are purely selecting them on their AI predictions. (Interview HR professional)

However, to generate predictions on candidates, the AI tool heavily relied on data from existing employees. Specifically, the algorithm needed a labelled dataset consisting of employee attributes and labels designating ‘high-performing employee’ or ‘low-performing employee’ to detect which attributes were important to select new candidates on. HR professionals, who benefitted from their relationships with employees, took on a leading role in data collection. They positioned themselves as advocates of the technology by convincing employees of its benefits and encouraging them to generate data for the tool. For instance, HR professionals personally invited employees to participate in the game- and video-based assessments and gave multiple presentations to workers’ councils to gain their approval for collecting sensitive employee data. As a result of these efforts, HR professionals managed to generate sufficient data on employee attributes to train the algorithm. In turn, HR professionals played a critical role in defining and selecting data for the algorithm by drawing on their existing expertise. For example, the HR team helped to define labels for the algorithm based on their knowledge of what is a holistic indicator of employee performance in the organization. The team also evaluated and refined attributes included in the predictive model based on their professional vision and values. For instance, the AI tool revealed that employees who valued a good ‘work-life balance’ underperformed in the company, meaning that candidates who shared this career value would be rejected by the algorithm. The HR professionals felt that this result was problematic, as it misaligned with their long-term vision of creating a sustainable workforce. Consequently, they decided that this attribute had to be removed from the dataset, so that ‘work-life balance’ was not used as a criterion in selection decisions. In sum, HR’s role shifted from a focus on selecting candidates to a focus on selecting data for the algorithm. Although HR professionals gave up a core judgment-oriented activity to the algorithm, they managed to draw on and incorporate some of their existing expertise in the process of supplying and selecting data for the AI tool.

Hiring algorithms  199

Guardian: From Expert Methods to AI Outputs The introduction of the AI tool also changed how HR professionals guarded fairness in the hiring process, by guiding senior managers with the help of AI outputs rather than expert methods. Traditionally, the HR team perceived it as their role to help senior managers in assessing candidates fairly. They offered unconscious bias training and developed a selection framework that listed relevant selection criteria. As part of a 45-minute briefing before every job interview, HR professionals trained managers in recognizing and suppressing bias and educated managers on the selection criteria. Based on their initial judgment of candidates, they also suggested interview questions to ask. For instance, when an HR professional noticed that a candidate lacked modesty in her motivation letter, she proposed that the manager ask: ‘How do you feel about someone constantly highlighting their successes?’ However, senior managers often experienced HR’s briefings and suggestions as lengthy and trivial. HR professionals frequently complained that managers came in late, skipped the briefing, or were ‘not even listening and doing something on their laptops instead’ (interview HR professional). With the AI tool, the HR team hoped to gain more interest from senior managers in their interview briefings and suggestions. Specifically, they decided to incorporate insights from the AI tool in the unconscious bias training to bring more credibility and data-driven evidence to their claims. For example, rather than educating managers on general biases that violated fairness in assessment, such as a ‘similarity bias’ or ‘confirmation bias’, HR professionals began to communicate past human flaws revealed by the AI tool. The algorithm had discovered, for instance, that senior managers tended to hire candidates who were eager to take risks, while high-performing employees were likely to be risk-averse. Senior managers were intrigued by such data-driven and detailed insights, expressing that ‘this is useful information that helps to challenge our own biases in the process’ (field notes). Moreover, HR professionals shifted their attention from the expert-based selection criteria to AI outputs on candidates. During the interview briefing, they showed candidates’ predicted fit with the company (i.e. displayed as a score from 0 to 100 per cent) and their dominant traits and skills (i.e. displayed in visual graphs). The team soon realized that such AI outputs were not taken up by managers unless being translated into concrete actions, requiring new analytical skills of the HR team. The AI manager, a newly established role for the project, helped the team to interpret and transform the AI outputs, by teaching them basic principles of statistics and designing simpler data visualizations. While at the start of the project an HR professional struggled and confessed that she was completely overwhelmed by the data, the same professional proudly shared several months later: I already made lots of education myself on analytics. I can already see… oh, this is a correlation! She [AI manager] is my teacher. (HR professional)

200  Research handbook on human resource management and disruptive technologies

Armed with their new analytical skills, HR professionals increasingly focused their briefings on the AI outputs, explaining to managers how to incorporate the insights into their assessment. For example, the HR team helped managers to formulate interview questions based on the traits and skills that the algorithm had identified as weak. They also encouraged managers to use AI outputs to solve disagreements about candidates. While some managers noted that ‘this [AI] visibility is affecting our scoring’ (field notes), they reasoned that there was no harm in relying on the AI outputs as they entailed ‘objective facts’. The HR team was thrilled that the AI outputs enabled them to steer managers’ activities in stronger ways than before, believing that this allowed them to better guard fairness in the hiring process. As a result of strongly engaging with the AI outputs, HR professionals managed to strengthen their grip on managers’ hiring practices. Mediator: From Interpersonal Disputes to Human–AI Disputes A final shift centred on HR’s role as a mediator in the hiring process, in which HR professionals were called upon to manage conflicts between the human and AI assessment of candidates. Traditionally, the team had mediated interpersonal disputes related to human assessment, where they drew on their knowledge of the organization’s policies and employment law to solve conflicts. In some instances, for example, HR professionals faced candidates who voiced complaints about discriminatory practices of senior managers, such as in the case of a candidate stating that ‘I noticed a preference of the assessor for people with the same nationality as he has’ (company documents). In other cases, HR professionals mediated conflicts between senior managers, such as when one manager accused another manager of ‘hiring the same profile over and over again’ (field notes). The use of the AI tool called for new forms of involvement of the HR team, this time including disputes around the AI assessment. Senior managers, for instance, willingly referred to HR professionals when they disagreed with the AI assessment of candidates. Similarly, the HR team was confronted with candidates who felt misjudged by the algorithm or found it humiliating to be rejected after a one-sided video interview without a sign of human involvement. The team actively began to resolve such disputes by reviewing cases of objection against the AI assessment. For example, some candidates reached out to the team and demanded to be assessed differently, as they strongly disagreed with the algorithm. Typically, the HR team proceeded by checking with the vendor whether a technical problem had occurred, as explained by an HR professional during a team meeting: So, we had a case where a guy completely disagreed with the AI results. Like, the person gets a score that does not reflect the understanding of himself. And then we discussed it with the vendor on the call and asked them to assess from their side if it was perhaps some mistake in the system, or maybe some issue with his computer. But they said that the candidate was not… he didn’t pay attention. So, it was no technical issue from their side. This is proven. (Field notes)

Hiring algorithms  201

Without evidence of a technical problem, the HR professionals decided to dismiss the candidates’ concerns. Instead, they focused their attention on justifying the AI assessment to candidates by referring to the tool as being ‘naturally unbiased’ and the complex inner workings of machine learning algorithms. In the observed instance where a technical problem did occur, HR professionals spent many hours ‘repairing’ faulty evaluations. This happened in the case of a software error, where the algorithm had mistakenly rejected multiple candidates. The HR professionals personally reached out to mistreated candidates and offered them the possibility to retake the game- and video-based assessments. In sum, HR professionals shifted their mediating role by including disputes around the AI assessment; activities that became in high demand once the tool began to assess candidates and influenced selection decisions. The team’s willingness to be a mediator in these sensitive situations placed them in a position of influence over senior managers and candidates, in which HR professionals had an important voice in settling conflicts between the human and AI assessment of candidates.

DISCUSSION The use of AI has been praised as an opportunity to strengthen the HR position by improving decision-making (Ulrich and Dulebohn, 2015; Budhwar et al., 2022; Vrontis et al., 2022), while others highlight that the same technology might undermine the HR position by marginalizing professional judgment (Angrave et al., 2016; Cheng and Hackett, 2021; Meijerink et al., 2021; Wiblen and Marler, 2021). In this chapter, we focus on what happens to the HR role in practice when professionals engage with AI tools. We find that the tool provided an occasion for HR professionals to reshape their roles in the hiring process. Rather than interpreting AI as threatening their expertise, HR professionals embraced it as an opportunity to elevate their position. However, in the process of engaging with the tool, HR’s activities increasingly centred on serving the needs of the algorithm. We conclude that engagement with AI in practice results in a dilemma for the HR profession: on the one hand, they may increase their influence over senior managers, but on the other hand, reduce the opportunity to develop and apply their unique expertise. Our findings build on prior work that has warned of a loss of human agency with the uptake of AI tools (Meijerink et al., 2021; Wiblen and Marler, 2021). We illustrate how HR professionals willingly gave up aspects of their traditional expert work and delegated it to the algorithm. As a result, HR professionals lost the authority and opportunity to apply their judgment on which candidates entered the organization, with the algorithm performing this form of expert work instead. We suggest that HR professionals can be lured into delegating decision-making authority to algorithms based on the strong outperforming claims associated with AI tools. For example, at MultiCo, the AI tool was positioned as being more efficient, rigorous and objective than human decision-makers in the hiring process. This implies that, rather than questioning or rejecting disruptive technologies (Greasley and Thomas, 2020; Wiblen and

202  Research handbook on human resource management and disruptive technologies

Marler, 2021), HR professionals can end up embracing AI tools in decision-making when being seduced by the tool’s glow. Our findings also indicate that HR professionals may gain legitimacy in the eyes of senior managers by taking up activities that support the AI tool. For example, MultiCo’s HR professionals helped to create and explain AI outputs to senior managers who valued these data points for being ‘objective facts’. In this way, HR professionals managed to gain influence over managers by steering their activities and decisions with help of the AI outputs. This observation resonates with prior work that showed that HR professionals strive to integrate data-driven tools as their outputs are appreciated by senior managers who speak a language of numbers (Greasley and Thomas, 2020; Ellmer and Reichel, 2021). Interestingly, HR professionals take up emerging tasks around these tools even though the tasks themselves do not exhibit enhanced epistemic value, but merely provide an allure of objectivity and legitimacy to the client that the profession serves. This chapter, therefore, suggests that HR’s quest for a more prominent role in the organization with AI is more ambiguous than commonly assumed (Ulrich and Dulebohn, 2015; Margherita, 2021; Budhwar et al., 2022; Vrontis et al., 2022). For example, while HR professionals expressed that the tool ‘empowered’ them by gaining voice in relation to senior managers, at the same time, this came at the expense of their unique expertise, training and judgment. To gain a seat at the table, HR may therefore find itself ‘mudding the boundaries of HRM expertise’ by taking out the human aspect of HRM and neglecting other key stakeholders (Wright, 2008: 1082). As a result, the HR profession risks becoming indistinguishable from rival groups, in which managers, data scientists or algorithms undertake activities traditionally reserved to HR.

CONCLUSION AND RECOMMENDATIONS FOR ACTION In conclusion, this chapter has highlighted that AI tools can reshape HR roles in ways that lead to a dilemma for HR professionals, in that they may on the one hand gain influence, while on the other hand lose the opportunity to develop and apply distinct expertise. Given the risk of losing unique HRM expertise, we call for reflexive HR professionals that routinely observe and reflect on the expertise, relationships and skills they draw on to perform their work. Rather than blindly accepting that AI ‘knows better’, HR professionals should understand what insights are gained and lost with the use of algorithms, and at what stages algorithms should be complemented with professional judgment. This allows HR professionals to see through the ‘hype’ and adopt a realistic approach to AI, in which they preserve and actively insert their expertise in the process, such as by defining AI’s objectives, aligning predictive models to the organizational context and providing a human touch to support and explain AI-based decisions. In turn, we recommend that HR professionals critically reflect on the ethical risks and responsibilities that arise with the uptake of AI in the workplace. Despite HR’s

Hiring algorithms  203

traditional role as ‘steward of the social contract’ between employees and the organization (Kochan, 2004: 133), this chapter shows how the belief in AI can make HR more closely aligned to the interests of management and vendors than the interests of employees. Given that the use of AI exposes employees to serious new risks related to privacy, profiling and surveillance, it is critical that HR professionals reflect on what the focus on AI implies for their traditional social contribution to the organization. Although this chapter is based on evidence from the specific case of hiring, we hope that these insights inform broader discussions on the implications of new AI tools for the roles and future of the HR profession.

NOTES 1. All names are pseudonyms. 2. For an extended description of the setting and data collection methods, see: Van den Broek et al. (2021).

REFERENCES Angrave, D., Charlwood, A., Kirkpatrick, I., Lawrence, M., and Stuart, M. (2016). HR and analytics: why HR is set to fail the big data challenge. Human Resource Management Journal, 26, 1–11. Beer, M. (1997). The transformation of the human resource function: resolving the tension between a traditional administrative and a new strategic role. Human Resource Management, 36(1), 49–56. Bondarouk, T., Parry, E., and Furtmueller, E. (2017). Electronic HRM: four decades of research on adoption and consequences. The International Journal of Human Resource Management, 28(1), 98–131. Budhwar, P., Malik, A., De Silva, M. T., and Thevisuthan, P. (2022). Artificial intelligence– challenges and opportunities for international HRM: a review and research agenda. The International Journal of Human Resource Management, 33(6), 1065–97. Burrell, J. (2016). How the machine ‘thinks’: understanding opacity in machine learning algorithms. Big Data & Society, 3(1), 1–12. Cheng, M. M. and Hackett, R. D. (2021). A critical review of algorithms in HRM: definition, theory, and practice. Human Resource Management Review, 31(1). Ellmer, M. and Reichel, A. (2021). Staying close to business: the role of epistemic alignment in rendering HR analytics outputs relevant to decision-makers. The International Journal of Human Resource Management, 32(12), 2622–42. Faraj, S., Pachidi, S., and Sayegh, K. (2018). Working and organizing in the age of the learning algorithm. Information and Organization, 28(1), 62–70. Greasley, K. and Thomas, P. (2020). HR analytics: the onto-epistemology and politics of metricised HRM. Human Resource Management Journal, 30(4), 494–507. Heizmann, H. and Fox, S. (2019). O Partner, Where Art Thou? A critical discursive analysis of HR managers’ struggle for legitimacy. The International Journal of Human Resource Management, 30(13), 2026–48. Houser, K. A. (2019). Can AI solve the diversity problem in the tech industry: mitigating noise and bias in employment decision-making. Stanford Technology Law Review, 22, 290–354.

204  Research handbook on human resource management and disruptive technologies

Kahneman, D., Sibony, O., and Sunstein, C. R. (2021). Noise: A Flaw in Human Judgment. William Collins Publishers. Kochan, T. A. (2004). Restoring trust in the human resource management profession. Asia Pacific Journal of Human Resources, 42(2), 132–46. Köchling, A. and Wehner, M. C. (2020). Discriminated by an algorithm: a systematic review of discrimination and fairness by algorithmic decision-making in the context of HR recruitment and HR development. Business Research, 13(3), 795–848. Kovach, K. A., Hughes, A. A., Fagan, P., and Maggitti, P. G. (2002). Administrative and strategic advantages of HRIS. Employment Relations Today, 29, 43–8. Langley, A. (1999). Strategies for theorizing from process data. Academy of Management Review, 24(4), 691–710. Lawler, E. E. and Mohrman, S. A. (2003). HR as a strategic partner: what does it take to make it happen? Human Resource Planning, 26(3), 15–29. Legge, K. (1995). Human Resource Management: Rhetorics and Realities. Macmillan. Margherita, A. (2021). Human resources analytics: a systematization of research topics and directions for future research. Human Resource Management Review, 100795. Marler, J. H. and Boudreau, J. W. (2017). An evidence-based review of HR Analytics. The International Journal of Human Resource Management, 28(1), 3–26. Marler, J. H. and Parry, E. (2016). Human resource management, strategic involvement and e-HRM technology. The International Journal of Human Resource Management, 27(19), 2233–53. Meijerink, J., Boons, M., Keegan, A., and Marler, J. (2021). Algorithmic human resource management: synthesizing developments and cross-disciplinary insights on digital HRM. The International Journal of Human Resource Management, 32(12), 2545–62. Orlikowski, W. J. (2000). Using technology and constituting structures: a practice lens for studying technology in organizations. Organization Science, 11(4), 404–28. Pessach, D., Singer, G., Avrahami, D., Chalutz Ben-Gal, H., Shmueli, E., and Ben-Gal, I. (2020). Employees recruitment: a prescriptive analytics approach via machine learning and mathematical programming. Decision Support Systems, 134, 113290. Rasmussen, T. and Ulrich, D. (2015). Learning from practice: how HR analytics avoids being a management fad. Organizational Dynamics, 44(3), 236–42. Sandholtz, K., Chung, D., and Waisberg, I. (2019). The double-edged sword of jurisdictional entrenchment: explaining human resources professionals’ failed strategic repositioning. Organization Science, 30(6):1349–67. Tambe, P., Cappelli, P., and Yakubovich, V. (2019). Artificial intelligence in human resources management: challenges and a path forward. California Management Review, 61(4), 15–42. Truss, C., Gratton, L., Hope-Hailey, V., Stiles, P., and Zaleska, J. (2002). Paying the piper: choice and constraint in changing HR functional roles. Human Resource Management Journal, 12(2), 39–63. Tursunbayeva, A., Pagliari, C., Di Lauro, S., and Antonelli, G. (2022). The ethics of people analytics: risks, opportunities and recommendations. Personnel Review, 51(3), 900–21. Ulrich, D. (1998). A new mandate for human resources. Harvard Business Review, 76, 124–35. Ulrich, D. and Brockbank, W. (2005). The HR Value Proposition. Harvard Business Press. Ulrich, D. and Dulebohn, J. H. (2015). Are we there yet? What’s next for HR? Human Resource Management Review, 25(2), 188–204. Van den Broek, E., Sergeeva, A., and Huysman, M. (2021). When the machine meets the expert: an ethnography of developing AI for hiring. MIS Quarterly, 45(3), 1557–80. Vrontis, D., Christofi, M., Pereira, V., Tarba, S., Makrides, A., and Trichina, E. (2022). Artificial intelligence, robotics, advanced technologies and human resource management: a systematic review. The International Journal of Human Resource Management, 33(6), 1–30.

Hiring algorithms  205

Wiblen, S. and Marler, J. H. (2021). Digitalised talent management and automated talent decisions: the implications for HR professionals. The International Journal of Human Resource Management, 32(12), 2592–621. Wright, C. (2008). Reinventing human resource management: business partners, internal consultants and the limits to professionalization. Human Relations, 61(8), 1063–86.

PART V DIGITAL DISRUPTION OF WORK PROCESSES

15. Engagement with disruptive technology: do digital generations matter?1 Frank Stegehuis and Tanya Bondarouk

INTRODUCTION According to the recent IT literature, IT has always been known for causing a rapid speed of change within businesses (Kaplan and Heinlein, 2019; Davison et al., 2019; Dittes et al., 2019; Cheng et al., 2020; Oberländer et al., 2020; Wang et al., 2020). But in addition to this, IT has institutionalized and is becoming an integral part of businesses. In this way, it is continuously altering the way employees do their work as new digital tools continue to emerge. Interconnectivity, being able to work regardless of place and time, is a new concept that reflects how IT is transforming businesses into a new era of work (Davison et al., 2019; Dittes et al., 2019; Kaplan and Heinlein, 2019; Cheng et al., 2020; Oberländer et al., 2020; Wang et al., 2020). Scholars provide several examples, stemming from empirical and conceptual studies, of the disruptive effects of technology. Wang et al. (2020) found that the usage of IT systems has a significant influence on the job satisfaction of employees, but that only a mere 9% of practitioners embrace improving the IT user’s experience. Cheng et al. (2020) argued that, while IT can bring convenience to employees, it also has negative influence through the frequent interruptions it can cause in one’s workday. Their study has shown the link between interruptions caused by IT and emotional exhaustion, which is a common precedent for job-related burnout. The examples of recent empirical studies mentioned above show that one may expect evolved routinization of IT-user interlacement. However, the reality is that the users are affected and disrupted by the digital tools that IT provides in their everyday job. It looks like historical developments do not demolish the disruptive effect of IT on users. These affections are conceptualized by some scholars in terms of ‘agency conflicts’ between the user and the technology. The agency theory, as applied to the IT usage, elaborates on how users enact with technology. They apply their ‘agency’, in other words goals and needs, onto a technological artifact. In this sense, they want to explain the technology and use it in order to achieve their goals and needs (Boudreau and Robey, 2005; Leonardi, 2013). Apart from users, the technological artifacts also have the ability to independently ‘constrain human agency once they are installed and left to operate’ (Boudreau and Robey, 2005, p. 4) through the limited set of options that they provide. Thus, both user and technological artifact enact with each other, often leading to consequences like the ones described in the previous paragraph. 207

208  Research handbook on human resource management and disruptive technologies

Introduction to Digital Natives and Digital Immigrants However, these consequences are unpredictable, due to the fact that IT usage and IT affection differs among individuals (Boudreau and Robey, 2005; Leonardi, 2010; Davison et al., 2019; Dittes et al., 2019; Cheng et al., 2020; Kesharwani, 2020; Wang et al., 2020). The differences among users of digital tools are depicted by some scholars through the division of a workforce into ‘digital natives’ and ‘digital immigrants’ (Dittes et al., 2019; Eginli and Isik, 2020; Kesharwani, 2020). Digital natives and digital immigrants are linked to the different generations of people that live and work in today’s society. According to Kesharwani (2020) and Eginli and Isik (2020), a digital native was born after the 1980s, and therefore exposed to digital technologies at a very early stage of his or her life. Digital immigrants were born before the 1980s, and thus before the rising importance of digital technologies in the workplace. Dittes et al. (2019) assume that digital natives extensively use digital technologies in their daily life and thus expect the same technologies at their work. Digital immigrants are described as not used to the new technologies and therefore very reluctant and critical towards them. Digital natives are also assumed to communicate differently with IT tools, by means of instant messages and online chats, whereas digital immigrants are described as using more traditional forms of online communications like e-mailing or calls. Lastly, it was assumed that digital natives use digital tools for networking activities whereas digital immigrants use it solely to increase their functionality (Kesharwani, 2020). The latter indicates that the age and thus generation of an end user has an effect on IT usage and affection. However, the work of Kesharwani (2020) and Dittes et al. (2019) is questioned by Eginli and Isik (2020) and Waycott et al. (2010) who argue that a number of synergies exist among generations. The work of Parry and Urwin (2017) questions if generational differences should be based on the age of an individual and suggests that more factors need to be included in order to uncover where true generational differences lie. Hence, differences could also lie within generations, rather than solely across them when they are divided by the age of individuals. Whether we agree with the age-related IT attributions that are described by the scholars and the debate on whether differences even exist among generations, we did find that researchers do see differences in working with digital tools that are related to user age. Hence, different digital generations of end users are assumingly affected by IT in different ways. An intriguing assumption is that if one considers the fact that generations of humans will always follow each other, together with the rapid and evolving development of IT within businesses, it is fair to assume that the digital natives of today could actually become the digital immigrants of tomorrow. Thus, the role of the end user generations and their relationship with digital tools needs a closer look if businesses want to avoid repetitive issues that emerge from the human-IT relationship and cope accordingly with the current trend of IT. However, the scholarly debate into digital generations and IT views technology to be unchangeable and rigid. For instance, Desouza et al. (2007) state that ‘The IT literature has mostly treated users as passive

Engagement with disruptive technology  209

consumers of technology’ (p. 205), implying that the user simply uses the technology based on its design and depicting the technology as a rigid artifact. In this chapter, we claim that it is time to explore possibilities for digital tools to adapt to the users. Therefore, the goal of this study is to explore the possibilities of IT becoming the adaptive agent in the human-IT interactions that occur at the workplace. To find out possibilities for this proposed ‘technological adaptivity’, our study was guided by the research question: What are user-generation characteristics of technological adaptivity in modern-day organizations? To address this question, in what follows, we briefly present the theory of agency and the literature on digital generations and IT to introduce starting insights. We then explain our research methodology and our present findings on characteristics of technological adaptivity towards different user generations. Adding to this, we add to the scholarly debate regarding digital generations and IT by exploring whether differences exist among them. Practitioners can use the implications from this work to improve the job satisfaction and productivity of their employees, because, after all, their employees are the end users of IT.

AGENCY THEORY, GENERATIONS AND USE OF IT The theory of agency elaborates on the enactment between end users and technological artifacts (Orlikowski, 1992; Boudreau and Robey, 2005; Cousins and Robey, 2005; Leonardi, 2010, 2011, 2013; De Boer and Slatman, 2018; Anaya, 2020; Hultin, 2020). Both the end user and the technological artifact (hereafter: digital tool) have a different perspective when it comes to their relationship. We start off with the end user perspective. Boudreau and Robey (2005) provide a good starting point, stating that ‘humans are free to enact with technology in different ways’ (p. 3). Cousins and Robey (2005) depict these different ways of enactment with technology further. The authors argue that end users may enact technological appliances as designers intended or they may improvise with technology to produce unintended patterns of use. Hence, end users use digital tools in differing and often unintended ways. Leonardi (2013) and Orlikowski (1992) explain these differing ways of usage by taking the end user’s specific goals and needs into account when he or she is interacting with a digital tool. The latter has caused the concept of ‘human agency’ to emerge among scholars (Orlikowski, 1992; Boudreau and Robey, 2005; Cousins and Robey, 2005; Leonardi, 2010, 2011, 2013; De Boer and Slatman, 2018; Anaya, 2020). We view human agency as the ability of a human being to set and realize goals. However, it is not something that is owned by a specific actor. Rather, it is the appliance by an actor of their goals or needs to a specific object or phenomenon (Orlikowski, 1992; Leonardi, 2011, 2013; De Boer and Slatman, 2018). As described by Leonardi (2013), people ‘Attribute their agency to equipment, machines, formulae and other various apparatus to explain the machinations of the universe through the imposition of causality’ (p. 62). Thus, in case of interactions with IT, human agency consists of how humans enact with technology to explain it and how they use it to

210  Research handbook on human resource management and disruptive technologies

achieve their goals and needs. An appliance of agency must therefore be seen as the ‘options of action’ that an end-user theorizes about when using technology, thereby also choosing if they appreciate it or not (Orlikowski, 1992; De Boer and Slatman, 2018). When end-users apply their own unique agency on technological artifacts, it could lead to them using the artifacts in ways that were not intended by the artifact’s designer (Orlikowski, 1992; De Boer and Slatman, 2018; Cousins and Robey, 2005). These unintended ways of usage result in a variety of effects depicted in a number of empirical studies that capture agency and technology usage within organizations. For example, Boudreau and Robey (2005) write that human interaction with technology results in two concepts, that of ‘inertia’ and ‘reinvention’. Basically, inertia describes humans avoiding the use of technology for various reasons like the novelty of it and how it isn’t their ‘used way of doing things’. Furthermore, workers also illustrate reinvention in which they do not use technology for its intended purpose, but instead work around it by using the system in an unusual, sometimes hazardous, manner. Thus, the end users of the technology have applied their agency, which has caused them to either not work with the artifact or work around the artifact’s intended purposes. The empirical work of Leonardi (2011) highlighted how the agencies of multiple crash-test engineers continuously led to the change of work routines and the functionalities of a digital tool. A new tool was implemented with the purpose of automating the crash-testing process and therefore improving the efficiency of the organization. The end users of the digital tool began applying their agencies, using it in a way that was consistent with their own goals. The engineers perceived the digital tool to be a constraining factor on their ‘standard routines’ and thus used it only for their own specific needs. Thus, the studies that apply the agency theory show that the specific goals and needs of the user of technology are the key determinant for the various consequences that emerge from the human-IT relationship. These consequences are often of a damaging nature to an organization because the workers do not ‘instantly adopt’ new technologies and their prescribed functionalities. Rather, the appliance of the user’s agency on technological artifacts is depicted as to why digital tools are used in an unintended and unanticipated way or not used at all. But, as mentioned before, these end users are all unique individuals who possess different goals and needs. That is why we now turn to the literature on generations and IT that provides more insights on this matter. Digi-generations and Agencies Table 15.1 provides a quick overview of the main differences between the two groups of technology end users (Kesharwani, 2020, p. 3). Whereas digital natives are seen as active end users and use the newest forms of technology (online chatting, creating online content), digital immigrants seem to use the more traditional forms of technology usage and are believed to show passive involvement. The question remains if digital natives, based on their early exposure to new technologies, adopt and work with these new technologies in a quicker fashion than the digital immigrants.

Engagement with disruptive technology  211

Table 15.1

Assumed differences between Digital Immigrants and Digital Natives in using IT

Basis

Digital Immigrant

Digital Native

Communication

E-mails

Online chats

Mobile Phone

Calls

Instant messages

Information Sharing

Limited and occasional sharing (very

Unlimited and frequent sharing (about daily

important things)

life happenings)

Blogging

To discuss thoughts with their peers; use as an To share personal thoughts publicly and use open discussion forum

blogging sites as diary

Usage Behavior

Single task: user of online content

Multitasking: creator of online content

Involvement level

Passive user; part of professional life

Active user; part of personal as well as

Primary use

To increase functionality

Networking: Interactivity

professional life

Source: Adapted from Kesharwani (2020).

A study by Kesharwani (2020) has shown that digital natives and digital immigrants do differ in terms of post-adoptive technology usage. Based on ‘sequential belief updating’, which represents the usage of technological artifacts in relation to past experiences and successes, and feedback mechanisms, it appears that digital natives show more continued usage behavior than digital immigrants. As argued by Kesharwani (2020), ‘Digital Natives are already using the technology themselves, while Digital Immigrants need a constant reminder to use it and more technology demonstration’ (p. 14). Both groups need to be trained differently based on technological skills. The study links the differences to a certain ‘digital inequality’, which points to an advantage position for the digital natives in terms of technological skills and experience. However, we assume that the goals and needs of these digital natives are more technologically oriented or supplemented than those of digital immigrants. After all, digital natives grow up with new technologies and use them more frequently than digital immigrants. Hence, we may assume that the digital natives are more comfortable with the new technologies which could shape their agencies to be more synergized with the digital tools in their work-environment (Waycott et al., 2010; Tilvawala et al., 2014; Eginli and Isik, 2020; Kesharwani, 2020). However, whereas both Kesharwani (2020) and Tilvawala et al. (2014) acknowledge a ‘clear divide’ between both groups in terms of adaptivity to new technologies, Eginli and Isik (2020) and Waycott et al. (2010) argue that this division is questionable. Their empirical studies show that a number of synergies exist between digital natives and digital immigrants. They argue that a better understanding about the perspectives of both groups is needed to understand the different forms of technology usage and interaction. Parry and Urwin (2017) add to the latter, arguing that a difference in generations should not be tied to the age of an individual. Rather, there are more factors that need to be uncovered. While it seems that scholars are arguing about whether digital natives and digital immigrants are really separable or not, we assume the different perspectives mentioned by Eginli and Isik (2020) and Waycott et al. (2010) to be differences in agency between the groups. Because

212  Research handbook on human resource management and disruptive technologies

both groups have experienced technologies differently, they appear to have different technological backgrounds. It is therefore arguable that their agencies (read: goals and needs) are shaped differently towards technologies at work. Thus, we assume that possible differences between generations are not related to age, but to goals. The latter indicates that differences possibly exist within generations rather than across generations when they are divided based on the age of an individual. Thus, if we view the human-IT relationship from the perspective of the user, both the agency theory and the theory on digital generations and IT indicate that differences exist between end users in terms of approaches, perspectives, goals and needs. Technological Agency Not only end users possess agency, because the technological artifact has its own form of agency as well. Leonardi (2013) defines this ‘technological agency’ as the ability to empower humans to act and to act independently of human agency ‘affording certain uses and actions’ (p. 70). Erofeeva (2019) further clarifies this ability by explaining that an object can make someone or something else say or do things throughout the options it provides them. For example, when end users perceive that an artifact offers no affordances for action, they instead experience that it constrains their ability to carry out their goals (Anaya, 2020). Hence, a technological artifact forces its end users to act in a certain way based on the options it provides them; this causes human agency to be constrained by this technological agency and causes technological artifacts to become ‘sociomaterial’. A sociomaterial artifact is co-shaped by the constant interaction between the user, who tries to achieve his or her goals, and technology, which provides a limited set of options for the user to choose from (Orlikowski, 2010; Leonardi, 2011, 2013; Erofeeva, 2019; Anaya, 2020). Thus, if we view the human-IT relationship from the perspective of the technology itself, it becomes apparent that the role of the technology is more disruptive than one may think: through the technological agency in the form of available options and affordances, the artifact constraints the human agency of its user. The constant interaction between the two agencies results in a technological artifact becoming sociomaterial and shapes it into a form that applies to its specific context. The shaping of a technology being sociomaterial gives an indication of the technological adaptivity that we propose in this chapter. However, it appears that, throughout the literature on agency, the user continues to be the sole initiator of adaptivity.

METHODOLOGY – ONLINE INTERVIEWS By interviewing both digi-generations, we aimed to check the key insights from the literature while collecting more insights on the existence of differences of interaction with digital tools between the two groups of end users. We also included software designers to interview them about possibilities of technology in terms of its functionalities and possibilities.

Engagement with disruptive technology  213

Table 15.2

Interviewee function and generation

Alias

Job profile

Generation

NURSE

Lactation consultant and

Digital Immigrant

POLICE

Advisor on capacity management for

premature-born baby nurse Digital Immigrant

a Police institution FINAD

Financial advisor for a large banking

Digital Native

firm SPEAKER

Public speaker for a governmental

Digital Native

organization SUPPLY

Stock and supply manager for a large

Digital Native

retail company APPMAN

IT application manager for a large tech Digital Native

ANALYTIC

Manager of the HR-Analytics

retail company Digital Immigrant

department of a large banking firm WEBDEV

Website developer and designer

BUSAPP

Developer of analytical IT applications Digital Native

Digital Native

SALESSUP

Sales support employee for a large

for Businesses Digital Immigrant

industrial company WPMAN

Workplace application manager for

Digital Native

a large tech retail company UNI-

Student-assistant for a Dutch

ASSIST

University

Digital Native

We randomly selected and invited the end user and the designer from any organization for an interview by means of an e-mail, telephone call or in-person approach. The selected 12 interviewees, their job profiles and their respective generation are displayed in Table 15.2. Data Analysis The interviews were recorded and fully transcribed and notes were taken during the conduction of them. The average time of an interview was 43.55 minutes and the transcripts had an average word count of 4602 words. The resulting transcripts were analyzed through the process of open coding to identifying characteristics of technological adaptivity. To provide the needed structure in the coding process, we used the model of Ollerenshaw and Creswell (2002). The ‘themes’ generated through this model will illustrate the ‘characteristics’ that appear to be existent as well as the findings on generational differences. The themes were generated through the iterative process of rereading and continuously filtering and grouping the retrieved insights from the interview transcripts. We undertook several steps within the data analysis to strengthen our arguments and to come to a proper conclusion (Table 15.3). Through

214  Research handbook on human resource management and disruptive technologies

Table 15.3

Stepwise visualization of analytical process

Step

Action

1

Conducting the interviews. During the conduction of the interviews, we were sensitive to reoccurring topics, demands and remarkable quotes and made notes of these.

2

While transcribing the interviews we were also sensitive to analyzing the reoccurring items. Notes from the interviews were compared with the transcripts to ensure that no valuable data was lost and the transcript was reread after it was finished.

3

After finishing all the transcripts, they were read twice before starting with the coding process to ensure that the themes and topics were clear.

4

The transcripts were coded within Atlas.ti. During the open coding process, we did not stick to already fabricated open codes. Instead, the transcript was carefully read and every section that contained relevant information for this research was coded.

5

129 open codes were generated. These open codes were first screened to find redundant codes.

6

We analyzed the remaining 118 codes and filtered the open codes down to those that showed clear dominance among the interviews or those that were found to be remarkable for this study. We put the 59 open codes that remained in 5 preconstructed code groups, which are: ‘End-user agencies’; ‘End-user interaction’; ‘End-user preferences’; ‘End-user perspectives’; and ‘IT-designer perspectives’.

this data analysis we were able to generate interesting findings. These are described in the next section.

FINDINGS End-User Agencies The goals and needs (read: agencies) of the two digi-generations were divided into personal or work-related. The digital immigrants mentioned goals and needs that were directly related to their job or function, for example: My goal is to help young parents with their baby. How they have to take care of it and especially how they can understand and take care of their baby in the first year. I want to have enough time at work to do it and not have too much of a workload. (NURSE) My goal is to advise the operational line within the police-organization, regarding the allocation of capacity versus work, as best as possible. With the support of a good office environment. (POLICE)

In comparison, when the same question was asked to a digital native he or she responded with goals and needs that were of a personal nature, as illustrated below:

Engagement with disruptive technology  215

My goal is to keep up with global developments. I don’t want to lag behind that is the purpose. I strongly favor a fun and social environment around me when I work in order to remain productive. (FINAD) My personal goal is to grow in leadership. I have been a specialist for many years and I now see that I have a need to become a better leader. (SUPPLY)

This distinct seperation between work-related and personal goals and needs did not come as a surprise, because we brought the element of age into the analysis. The digital immigrants were all of a more senior age. Whereas digital immigrants had been working for a significant amount of time, the digital natives were at the start of their careers and thus their goals were more personally oriented. However, we observed that the digi-generations in the sample did not feel a negative disturbance of digital tools on their agencies. Both viewed IT as an enforcing element regarding one’s goals and needs: The urgency to keep up with the developments has become bigger for me due to these IT applications. They show me what I have to prepare for and what is possible. (FINAD) I think that the because of IT applications chances and oppertunities are becoming visable. They make things measurable and you can see where you need to develop. Thus, if I want to develop myself I use them and if I want to use data they strenghten that goal as well. (SPEAKER)

Hence, the two digi-generations did differ in terms of agency but the agency conflicts that we assumed to orginate from these differing agencies were not apparent. On the contrary, both digital natives and digital immigrants felt that digital tools in their workplace enforced their goals. We did not expect this synergy between the technological and human agencies. Moving on, we took a closer look at the possible differences in the interaction with digital tools between the two digi-generations. End-User Interactions: Positive Experiences Overall, the digi-generations in our study did not show significant differences regarding their interactions with digital tools. Both digital natives and digital immigrants mentioned that they experienced the overall influence of digital tools within their work as positive: You do not only see it with our stock-taking tasks but also with registration of certain sale loops in our systems. Everything is just supported better by IT-systems. (SUPPLY) I have never seen it as a threat. On the contrary, I have always embraced it because it helps you in so many cases. I never found it annoying. (SALESSUP) I see it as very useful, not neglecting the fact that me and most of my fellow colleagues saw it as a very large step. (NURSE)

216  Research handbook on human resource management and disruptive technologies

Hence, both digital natives and digital immigrants viewed digital tools as something positive in their work environment. As we showed earlier in this chapter, the literature assumes that digital natives would be more comfortable with digital tools than digital immigrants. However, we observed that digital immigrants also had a positive view on them. Secondly, when we narrow the scope down to the specific interaction with digital tools, both digi-generations mentioned that technologies allowed for a more efficient and effective way of working: Well, in the past when multiple people were working in one document you continiously had to save and send the file back and forth which costs a lot of time. Now we just work together in one document and we can negiote while doing so. (ANALYTIC) Well, if we talk about the application that I use for my university-job I can use it to reach out to all the students at once. I can for instance post certain messages on a discussion board and they will all get notified. (UNI-ASSIST)

Thus, all of the end users within our sample appeared to have a positive view on the presence of digital tools in their work environment and they experienced them as a supportive element during their workdays. We found this interesting because it was assumed that digital natives would be more optimistic about digital tools than digital immigrants. However, we observed a shared optimism between the two digi-generations. Both digi-generations made active use of digital tools either because they had to or because they wanted it themselves. It was assumed that digital natives would show more active involvement with digital tools, but this does not seem to be the case within our sample. Apart from these positive experiences, the end-user interactions also lead to negative experiences regarding digital tools. End-User Interactions: Negative Experiences We observed that both digi-generations mostly experienced the same frustrations when working with digital tools, apart from one differing topic. First, we found that a failing digital tool was a common cause of frustrations: I have for instance experienced that I had a group of 30 people in a conference room, but the beamer and powerpoint presentation would not work. You know, that I had a presentation but the technology failed to work. (NURSE) Well it is at that moment that when there is a malfunctioning in your IT infrastructure that you know that you can not do anything anymore. You just sit in your chair and wait till the problem is solved. (UNI-ASSIST) Well if I can not perform my tasks anymore because an IT application is not working. That is mostly due to the internet connection though. (POLICE)

Thus, in most situations a failing digital tool left an end-user unable to perform his or her tasks, which blocked them in their productivity. The police advisor already introduced the second negative experience: a failing internet connection. We heard

Engagement with disruptive technology  217

voices from both digi-generations that internet connections were also a common cause of their negative experiences with digital tools. Lastly, we observed that both end user groups found it very frustrating when they had to take unnecessary steps while using a digital tool: I find the expert-environment to be nice because I can approach several platforms and retrieve data without having to do any complicated stuff. (ANALYTIC) We had an EPD and I had to log in via the browsers. Then I got a text message with a verification code on my phone before I could get in the application. And with another application I had to use a Citrix environment which was not available on every computer because you needed a certain connection. That was quite cumbersome. (UNI-ASSIST)

Hence, end users from both digi-generations mentioned that having to take a lot of, in their eyes, unnecessary steps was also felt as a negative aspect of digital tools. We did not observe a dominance in difficulties that we assumed to be existent within the digital immigrant group. Learning and Fluency End users of digital tools, especially digital immigrants, either learned how to use them in their everyday job or had fluency with them. We observed that digital natives had indeed ‘grown up’ with digital tools as mentioned in the literature. They therefore already possessed adequate experience and knowledge. We observed that they experience less difficulty with digital tools because of this prior knowledge: I think that, when I compare it with my older colleagues, my young age and IT experience makes that I can use and adopt the digital tools much quicker. (FINAD) Well, If I want a certain application for the organization I have to consider that it is also suitable for my colleagues that are like 60 years and older. If I do it for my generation I can make it much more complex. (SPEAKER)

Secondly, we assumed that digital immigrants would have more difficulty with digital tools because they ‘missed’ growing up with them. However, we observed that this was not the case. The digital immigrants voiced that they had either learned how to use digital tools or that they had fluency with IT in general causing them to have adequate knowledge of digital tools. They therefore did not ‘miss the boat’ by not growing up with digital tools. In fact, some of the digital immigrants mentioned that because of their fluency towards the new IT-technologies, they did not miss any of the developments that took place when IT was still an upcoming phenomenon: In 1984 the police retrieved the first computers in a project. I was one of the first to be interested in that and was appointed as an instructor for that project. (POLICE)

218  Research handbook on human resource management and disruptive technologies

When I was still studying people started with programming in all kinds of programmer language. I was interested in the usage of a certain device and what you could do with it. I saw it as something very good, not a threat or anything. (SALESSUP)

Third, apart from having fluency and/or experience with digital tools both digi-generations frequently mentioned that they do not face a lot of difficulty using them because they learned how to use them: Yes everything is going fine now, it could be a bit more simple but it works because I now understand how it works. (NURSE) I did not know how the system functioned at the time, but that is not a big problem because everyone has that at a certain point. (UNI-ASSIST)

Thus, it appears that end users from both digi-generations did not face a lot of difficulties using digital tools in their work routine. We assumed that especially the digital immigrants would face more issues with the usage of digital tools than the digital natives, but we observed that the experienced difficulties with digital tools were not distinguishable between the digi-generations of our sample. Above all, we found that the presence of adequate knowledge and experience with digital tools, either through fluency or learning, had a positive effect on user-technology interaction. A set of demands originated from the end-user interactions with digital tools. End-User Demands Almost all of the uncovered demands were applicable to both digi-generations within the sample, apart from two distinctive ones. Firstly, both end user groups mentioned that they demanded ‘accessibility and performance’ from digital tools, which in simple terms meant that they ‘worked’. We saw this as a direct response to the difficulties regarding system malfunctions and internet connections that originated from the end-user interactions: It actually became like a commodity. You just expect it to work and if it works than you do not hear anyone complaining. (ANALYTIC)

The end users simply wanted digital tools they could rely on and that did not block them in their work routine in order to work fast and efficiently. Secondly, we found that both end user groups demanded synergy and overlap between the digital tools in their work environment: In the first place that those systems can communicate with each other. That when I write something down in one system it can get adopted by another system. (POLICE)

Third, both digi-generations also demanded that digital tools communicated with them. They mentioned examples like pop-ups that reminded them of upcoming tasks,

Engagement with disruptive technology  219

notifications that certain actions contained errors which needed to be solved and other actions that made the digital tool a type of ‘virtual-assistant’ to its end-user: Yes, for example that when I drive my car and I am not paying attention I get an alert that I am crossing the line across the road. He therefore gives me a signal that I need to pay attention. I think that this kind of technology should be much more incorporated, feedback from a certain application. (SALESSUP)

Lastly, both digi-generations described the preference for easy-to-use, clear and self-explanatory digital tools. Throughout the data collection and analysis we found that these aspects were actually part of an umbrella term, ‘intuitiveness’. We chose to use intuitiveness to describe these aspects because both end users and IT developers used this term frequently. In fact, it proved to be a concept that was closely tied to our proposed technological adaptivity: So, if you do not use a certain system that often it has to be entirely self-explanatory how you get to something that you need. In expert systems it can be quite a puzzle to get to where you want to go. (ANALYTIC) I found it a very user-friendly application and I think that is very important. (UNI-ASSIST)

Apart from these similar demands, we also uncovered demands that were applicable to only one digi-generation. Digital natives mentioned the importance of hardware when talking about their preferences: But also, the quality of IT. I have a work phone here that you can get for under a 100 euros. If I open the NOS-app it already malfunctions. I have to get my new iPhone in order to work. (SPEAKER)

The digital immigrants also had a demand that only accounted for their digi-generation. They mentioned that they preferred a system that had an aspect of intelligence within it: Well in terms of supporting, you also have that assistant that Google has developed. It is constantly learning, and I can imagine that it is going to automatically do certain tasks for you. (POLICE) Like in terms of intelligence of an IT application. That it sends you personalized messages, it reads your agenda and messages you that you have been typing for too long. It is possible and it is going to be upcoming, I am sure of that. (ANALYTIC)

Hence, the digital immigrants demanded that digital tools possessed intelligence and therefore assisted them in a more personalized matter. This observation was notable because intelligent digital tools should have been of more interest to the digital natives, based on our observation and assumption that they expect more from digital tools. However, they did not mention aspects like these while the digital immigrants did.

220  Research handbook on human resource management and disruptive technologies

DISCUSSION The central question that we addressed in this study was: ‘What are characteristics of technological adaptivity towards different user-generations in modern-day organizations?’ We asked this question because of the rising importance of IT technologies within organizations and the agency conflicts that occur from the relationship with their end users. We concluded that the presumed generational differences did not exist among the digi-generations in our study. We observed a shared optimism about digital tools among the digi-generations. Furthermore, we observed no significant issues with digital tools as well as a similarity in demands and perspectives. Whereas the literature on digital generations and IT mentioned that this optimism would be more apparent within the digital native generation and that these natives would be more competent with the use of digital tools, we found that the digital immigrants were similar to the natives. We tied this finding to the concept of learning and fluency, because both digi-generations had either learned how to use digital tools over time or already had adequate knowledge of IT in general either through growing up with them or having fluency towards them. Hence, we found that difficulties with digital tools were mainly applicable to general issues like internet connections and failing IT systems but not to the specific competency of the end user. The latter meant that the agency conflicts that we assumed to originate from the difference in technological skills based on generations of end users were not apparent. Thus, synergies do appear to exist between the two digi-generations as was also mentioned within the studies of Waycott et al. (2010) and Eginli and Isik (2020). Therefore, our study adds the concept of learning and fluency to the existent theories on digital generations and IT, which explains these similarities. We conclude that instead of age, the agencies of individuals (read: goals and needs) is the factor that explains differences between generations in IT usage, which is in synergy with the similar claims of Parry and Urwin (2017). Because of the differing experiences of individuals within and across generations of human beings, they each strive for different goals, which also affect their technology usage and affection. Agency conflicts served as the basis for the feedback that caused technology to become more adaptive towards its user. We saw the end-user input characteristic of adaptive technology as the product of the human-technology relationship that is created through the constant clashing of human and technological agencies. After all, we aimed to explore if the digital tools themselves could be more adaptive and therefore could avoid the agency conflicts that cause a loss in productivity and satisfaction in the first place. Lastly, we identified a couple of restricting factors as well. Organizations and their financial intentions proved to be the most restricting on the adaptivity of a digital tool. Moreover, the size of an organization influenced how adaptive the tool in question could be, because a large organization could not afford to pay for a tool that was tailored to all of its individual departments. This caused the digital tool to be developed based on the generic need of the entire organization, which reduces its adaptivity. Apart from these organizational factors, we also observed technical boundaries. These boundaries simply implied that not everything can be ‘coded’ or crafted in

Engagement with disruptive technology  221

a digital tool by an IT professional. To conclude, the interpersonal differences among end users made incorporating every specific need of an individual impossible. As one can see, the end users played an active role through their end-user input that influenced the IT design. As we mentioned earlier, this process was similar to the concept of imbrication, wherein the agency conflicts that arose through the interaction between end users and digital tools caused digital tools to become more adaptive. These combined characteristics influenced the IT design in such a way that the digital tools within work environments were already experienced by the end users as being tailored towards them and thus were causing technological adaptivity to occur. However, the IT design process was also influenced by a number of restraining factors that influenced the possibilities within IT design. This caused the technological adaptivity to not be fully optimized because the IT design was limited in its possibilities. Finally, the interaction between IT design and technological adaptivity is seen in this chapter as a reciprocal process. The adaptive characteristics and restrictions cause adaptive digital tools to originate from IT design, whereas the feedback from the resulting technological adaptivity flows back towards IT design and causes further improvement under the same influence of those adaptive characteristics and restrictions.

CONCLUSION AND RECOMMENDATIONS FOR ACTION It appeared that a distinct difference between the end users of technology based on their generation did not exist, but that possibilities for digital tools to adapt to its end users were available and already in play. We found that the digi-generations of our sample were not separable from each other, apart from minor anomalies, which affirmed that a distinct difference between the digital generations of IT-users is questionable. We uncovered that the differences between the digi-generations became diminishable through learning and fluency and that agency theory actually explains differences in technology usage among and within generations because of a difference in goals between individuals rather than a difference in age. We therefore added to the question of scholars to further investigate differences and similarities between the generations of IT users. Furthermore, we uncovered that three characteristics, being the incorporation of end-user input, the importance of user experience within the IT sector and the expected adaptive trend within that sector, are driving digital tools to become increasingly adaptive towards their end users. The incorporation of the end-user input within IT design proved to be a confirmation of the agency theory, wherein human and technological agencies clash and therefore cause digital tools to adjust based on their feedback. However, we also uncovered that digital tools are being tailored towards the end user before they are implemented. The latter meant that digital tools were not as rigid as we assumed them to be when they were introduced to a workforce and we observed that this will continue to evolve because of the increasing importance of user experience within the IT sector and the fact that this adaptive trend will likely continue in the future. The latter implied that agency

222  Research handbook on human resource management and disruptive technologies

conflicts and thus losses in productivity and work satisfaction could reduce because we assume that the technological agency of digital tools will become increasingly synergized with the human agency of its end users as this adaptive trend will continue. The presence of adaptive technological artifacts proved to be a phenomenon that was unaccounted for in the literature on agency and thus detaches from the original views that depicted technological artifacts as rigid and limited in their available options. It would be interesting to see where this novel view could take the agency theory in the future.

NOTE 1.

This chapter is based on Stegehuis, F.J. (2021). Digital Immigrants and Digital Natives: an explorative study into the adaptivity of technology. Unpublished thesis, University of Twente, The Netherlands.

REFERENCES Anaya, L.A. (2020). Realizing the benefits from enterprise information systems: a sociomaterial perspective. Procedia Technology, 9, 473–9. Boudreau, M.C. and Robey, D. (2005). Enacting integrated information technology: a human agency perspective. Organization Science, 16(1), 3–18. Cheng, X., Bao, Y. and Zarifis, A. (2020). Investigating the impact of IT-mediated information interruption on emotional exhaustion in the workplace. Information Processing & Management, 57(6), 1–16. Cousins, K.C. and Robey, D. (2005). Human agency in a wireless world: patterns of technology use in nomadic computing environments. Information and Organization, 15(2), 151–80. Davison, R., Ou, C. and Ng, E. (2019). Inadequate information systems and organizational citizenship behaviour. Information and Management, 57(6), 103240. De Boer, M. and Slatman, J. (2018). The mediated breast: technology, agency, and breast cancer. Human Studies, 41, 275–92. Desouza, K.C., Awazu, Y. and Ramaprasad, A. (2007). Modifications and innovations to technology artifacts. Technovation, 27(4), 204–20. Dittes, S., Richter, S., Richter, A. and Smolnik, S. (2019). Towards the workplace of the future: how organizations can facilitate digital work. Business Horizons, 62(5), 649–66. Eginli, A. and Isik, S. (2020). Generational differences in a digital age, a research on technology experiences of generations. International Journal of Scientific & Technology Research, 9(2), 3150–4. Erofeeva, A. (2019). On multiple agencies: when do things matter? Information, Communication & Society, 22(5), 590–604. Hultin, L. (2020). On becoming a sociomaterial researcher: exploring epistemological practices grounded in a relational, performative ontology. Information and Organization, 29(2), 91–104. Kaplan, A. and Heinlein, M. (2019). Digital transformation and disruption: on big data, blockchain, artificial intelligence, and other things. Business Horizons, 62(6), 679–81. Kesharwani, A. (2020). Do (how) digital natives adopt a new technology differently than digital immigrants? A longitudinal study. Information & Management, 57(2), 1–16.

Engagement with disruptive technology  223

Leonardi, P.M. (2010). Digital materiality? How artifacts without matter, matter. First Monday,  15(6). https://​doi​.org/​10​.5210/​fm​.v15i6​.3036. Leonardi, P.M. (2011). When flexible routines meet flexible technologies: affordance, constraint and the imbrication of human and material agencies. MIS Quarterly, 35(1), 147–67. Leonardi, P.M. (2013). Theoretical foundations for the study of sociomateriality. Information and Organization, 23(2), 59–76. Oberländer, M., Beinicke, A. and Bipp, T. (2020). Digital competencies: a review of the literature and applications in the workplace. Computers & Education, 146, 103752. Ollerenshaw, J. A. and Creswell, J. W. (2002). Narrative research: A comparison of two restorying data analysis approaches. Qualitative Inquiry, 8(3), 329–47. Orlikowski, W.J. (1992). The duality of technology: rethinking the concept of technology in organizations. Organization Science, 3(3), 398–427. Orlikowski, W. J. (2010). The sociomateriality of organisational life: considering technology in management research. Cambridge Journal of Economics, 34(1), 125–41. Parry, E. and Urwin, P. (2017). The evidence base for generational differences: where do we go from here? Work, Aging and Retirement, 3(2), 140–8. Tilvawala, K., Myers, M. and Sundaram, D. (2014). Design of organizational ubiquitous information systems: a framework for digital natives and digital immigrant end-users. Proceedings of the Pacific Asia Conference on Information Systems 2014, 132. Wang, W., Wang, Y., Zhang, Y. and Ma, J. (2020). Spillover of workplace IT satisfaction onto job satisfaction: the roles of job fit and professional fit. International Journal of Information Management 50, 341–52. Waycott, J., Bennett, S., Kennedy, G., Dalgarno, B. and Gray, K. (2010). Digital divides? Student and staff perceptions of information and communication technologies. Computers & Education, 54(4), 1202–11.

16. Artificial intelligence as a colleague: towards the workplace coexistence of people and artificial intelligence Violetta Khoreva and Katja Einola

INTRODUCTION Artificial intelligence (AI) is a modern-day buzzword. It is also a topic of ongoing debate, a source of endless possibilities and an essential element of disruptive technologies constituting the process of digitalization that is radically transforming our working life (Haenlein and Kaplan, 2019; Huang et al., 2019; Snell and Morris, 2021). The expansion of AI, big data and ‘smart machines’ also has its critics. Zuboff (1988, 2019) warns against ‘surveillance capitalism’ that relies on a global architecture of computer mediation and digital networks. This architecture is not only shifting power from governments to social media companies such as Google and Meta, but also triggering feelings of disorientation and fear of job losses amongst both blue- and white-collar workers (Lanier, 2014). In organizational practice, AI is an example of a disruptive technology that is increasingly becoming a part of our everyday life, in a similar way that the internet and social media did in the past. But there is an important difference. Unlike other technologies, AI solutions are designed to execute increasingly complex tasks previously conducted by humans. Many of these solutions go beyond the automation of simple tasks and include modules for machine learning that are not transparent to people using AI. Hence, the functionality of an AI solution goes beyond what a traditional technology platform or a digital tool can deliver, and the work the machine does and the human does is increasingly intertwined. In our view, the interesting question here is not so much whether or what kind of role AI plays in organizations and their human resource practices; rather, we find it fascinating to advance our understanding of how people and AI evolve and learn to coexist (Haenlein and Kaplan, 2019, p. 9). As this technology develops, humans learn to work with it and the technology becomes smarter. These two processes are tightly intertwined. This chapter is drawn on our recent empirical article (Einola and Khoreva, 2022) and looks to convey the realities of human/AI coexistence in real-world organizational settings. In studying the coexistence of people and AI in organizations, the paradox theory perspective (Smith and Lewis, 2011; Schad et al., 2016) offers a vantage point from where we can observe the dynamic interplay and complexity. Following the main idea of the article, we advocate that the coexistence of people and AI should be recognized as a new organizational phenomenon and a topic of study 224

Artificial intelligence as a colleague  225

which describes how both people and technology are inevitably intertwined and coexist, rather than thinking of either people or technology, while not forgetting the dynamic and ever-changing nature of this coexistence. What a machine can do today is not the same as its capabilities tomorrow. People are in charge and need to think proactively about how this coexistence evolves over time. We project an idea of a peaceful coexistence between people and AI, in part to counteract the threatening images the sci-fi movie industry has been so good at creating in films such as I, Robot, Star Wars and Terminator, and in part to bring nuance to the one-sided rhetoric in which AI is predicted to take over human jobs, making people redundant (see Fleming, 2019, for a critique), or to threaten our safety, privacy and even our civilization (Lindebaum et al., 2020).

WHAT IS ARTIFICIAL INTELLIGENCE IN AN ORGANIZATIONAL SETTING? Most research on AI has its roots in the fields of technology studies and computer science. The investigation of AI in organization studies in general and HRM in particular is much more recent and under-developed. Nilsson (2009) adopting a philosophical and historical approach, calls AI ‘a quest’. We find this characterization also fitting for organizations – we are still very much in search of what AI is in modern organizations. Is it a new kind of organizational actor with agency or simply a machine? Is it friend or foe? What will it become in the future? To avoid terminological confusion, we build on long-standing discussions among scholars, and clarify how we approach the phenomena of digitalization and AI. The term digitalization has been coined to describe a manifold sociotechnical phenomenon. It is a fundamental process of adopting and applying disruptive technologies at different levels (Haenlein and Kaplan, 2019). These technologies include, but are not limited to: AI, data analytics, robotics, digital platforms, digital twin, social media, digital traces, blockchain, and 3D printing (Einola and Khoreva, 2023). That is to say, AI is one of these artefacts, which together represent the process of digitalization. Raisch and Krakowski (2021, p. 192), following Nilsson (1971), define AI as ‘machines performing cognitive functions, which have traditionally been associated with human minds, such as learning, interacting, and problem solving’. Kaplan and Haenlein (2019, p. 15) refer to AI as ‘a system’s ability to interpret external data correctly, to learn from such data, and to use those learnings to achieve specific goals and tasks through flexible adaptation’. AI can also be considered shorthand for ‘a set of complex algorithms for data collection and analysis that makes sophisticated predictions and evaluations possible; it is capable of interacting with the environment and simulates or even exceeds human intelligence’ (Glikson and Woolley, 2020). Taken together, these definitions suggest that AI is flexible, self-adapting and human-like (or exceeds human capabilities); it takes over human tasks, engages in learning, is set to achieve specific goals and is capable of making predictions and evaluations and of interacting with the environment (Einola and Khoreva, 2023).

226  Research handbook on human resource management and disruptive technologies

While these statements are broadly true, they only apply to specific instances, tasks and functionalities. There is no generic AI capable of performing all these functions at once. Therefore, in the management literature, AI is conceptualized as having a twofold purpose: to take over simple jobs or routine tasks from humans (commonly referred to as ‘automation’) and to assist humans in more complex tasks, such as data analysis and decision-making (commonly referred to as ‘augmentation’) (Raisch and Krakowski, 2021). These AI-enabled automation and augmentation solutions (hereinafter ‘AI solutions’) are local, task-specific and captured by algorithms; defined by some people (ideally those who know the task) and coded by others (who know how to code) into a language the specific AI solution can ‘understand’. Once this code is ready, the AI is tested by its users and introduced into the workflow. Here, frequent adjustments of both the code and the organizational work processes are required to get the solution ‘right’. This is where coexistence between people and AI is put to the test: it can be peaceful when the introduction of AI goes smoothly, and conflictual when it causes problems and does not work as prescribed or promised, triggering stress, frustration and loss of employee time and motivation.

COEXISTENCE OF PEOPLE AND ARTIFICIAL INTELLIGENCE Because we cannot quite talk about relations between thinking, feeling people and inert AI, we prefer the notion of coexistence better to understand the phenomenon of AI in organizational settings in its early stages of development (Einola and Khoreva, 2023). Coexistence is a neutral, open concept that can capture a variety of relations (Kriesberg, 1998). It refers to the conditions that serve as the fundamental prerequisites for the evolution of advanced harmonious relations (Rothstein, 1999; Whittaker, 1999; Bar-Tal and Bennink, 2004). It denotes recognizing the right of each group to exist peacefully with its unique features, and the acceptance that any differences need to be dealt with nonviolently. The promise of coexistence is that it provides a springboard to stronger, more respectful intergroup relations in the future (Einola and Khoreva, 2023). Coexistence is not a commonly used concept among organizational scholars, and, therefore, compared with other concepts describing positive relations, is seldom used in this context (Weiner, 1998). This is due to the vague and indistinct nature of the term, and because it pertains only to minimally positive intergroup relations (Weiner, 1998). Hence, coexistence is an open concept that ‘leaves a great deal of room for various forms of relations’ (Kriesberg, 1998, p. 183). In our view, assuming positive relations between groups from the outset is wishful thinking, and does not take organizational complexity and human nature into consideration. To the best of our knowledge, today’s techno-hype and optimism, as well as techno-doom talk and pessimism, are not empirically rooted in studying actual organizational practice and HRM. Thus, many popular books and well-cited conceptual articles are

Artificial intelligence as a colleague  227

somewhat unhelpful if we are trying to understand the everyday challenges encountered by people working with these technologies in today’s organizational settings. Even though coexistence of people and AI is different from coexistence between people, we need a suitable vocabulary to discuss AI and other disruptive technologies as they become more integrated in organizations, and increasingly interactive and embedded in workflows involving both people and machines. The idea of assigning agency of some sort to inert things is not new (Latour, 1992), nor without controversy (Sayes, 2014). Moreover, it is still not clear how to treat nonhuman entities in social analysis. However, we can no longer neglect the emergence of AI as a new type of ‘coworker’, as AI-based robots, bots, virtual assistants, algorithms with machine learning modules and the like become more common and widely used, and as people assign human qualities to them (Mori, 1970). Thus, we suggest that to guide our inquiry and research, the term coexistence, traditionally used to describe relations between people, should be extended to refer also to interaction between people and AI. We recognize at the same time that these relations are not (and cannot be) based on ‘mutually shared foundations’, for the obvious reason that AI has no human qualities, such as emotions, sensorial experiences, intentionality, morality and so on (Moser et al., 2022). The otherwise criticized quality of openness, and the lack of direction of the concept ‘coexistence’, seems particularly appropriate to describe interactions between people and AI, a relatively new phenomenon we still struggle to understand. We thus define the coexistence of people and AI in organizational settings as ‘organizational members interacting with AI solutions, including any kind of contacts and bonds between people and AI generating beliefs, attitudes, emotions, and behavioural patterns, once the AI solution is implemented and as it evolves over time’ (Einola and Khoreva, 2023, p. 119). This definition emphasizes that human/AI coexistence triggers beliefs, attitudes and emotions in organizational members. It also highlights the temporal nature of the coexistence, in that it keeps evolving as AI solutions become more sophisticated and people learn to use them in different ways.

THE PARADOX THEORY PERSPECTIVE There is a tendency to divide AI applications into two types of solution: 1) more mundane automation solutions that replace humans with machines, and 2) more advanced augmentation solutions in which humans work in tandem with AI. The nature of the task, the workflow and the fit of the solution determine whether organizations may opt for one or the other solution for a given task at a given point in time. For example, if organizations opt for automation solutions (presently most suited to well-scripted, routinized tasks such as invoicing and order handling), humans hand the task over to a machine, with little or no further involvement on their part. In contrast, augmentation solutions imply continued close interaction between humans and AI. Here, the key point is that only relatively routine and well-structured tasks can be automated, whilst more complex and ambiguous tasks need to be addressed through

228  Research handbook on human resource management and disruptive technologies

augmentation (Raisch and Krakowski, 2021). This trade-off perspective, where organizations opt for one solution or the other, needs however to be more nuanced. The paradox theory perspective warns us against this type of either/or thinking based on trade-offs that may not adequately represent reality (Smith and Lewis, 2011). Adopting a paradox lens allows organizations to capture not only the contradictions but also the interdependencies between automation and augmentation. The essence here is that the two elements are both contradictory and interdependent, thus forming a persistent tension (Schad et al., 2016), as they rely on competing logics with different organizational demands (Krakowski and Raisch, 2021). Automation and augmentation are contradictory because organizations may choose either one or the other type of AI solution at a specific point in time for a specific task. These solutions are also interdependent, because they are integrated into a workplace ecosystem and workflows where they coexist not only with each other and humans in different occupational roles, but also with other technologies. The tension is further reinforced by purely human factors. Some organizational actors prefer augmentation solutions (e.g. managers at risk of losing their job to automation) while others prioritize automation (e.g. owners interested in efficiencies) (Krakowski and Raisch, 2021). If temporal (from one point in time to another) and spatial (from one to multiple tasks or from one location to another) aspects are considered, the complexity and the paradoxical nature of automation/augmentation are further amplified. Organizations may opt for one or the other solution at a given point in time, which mitigates the underlying tension temporarily but fails to resolve it. A task that initially requires humans working intensively with AI may eventually become an automated task requiring little or no human intervention as learning and further development occurs. It is also possible that what appeared at first to be a simple automation task requires tight human/AI cooperation, due to neglected or unexpected complexities involved in making the new process work. Environmental factors or technology changes may trigger further changes requiring adjustment in both AI solutions and human/AI coexistence.

COEXISTENCE OF PEOPLE AND ARTIFICIAL INTELLIGENCE FROM THE PARADOX THEORY PERSPECTIVE The complexity that characterizes AI is even greater if we consider that humans and AI solutions of any type are not placed in separate worlds but necessarily coexist. So, any inherent tensions are not purely between types of AI but inevitably involve people (Einola and Khoreva, 2023). From an organizational perspective, it is easier to implement an AI-enabled automation solution than an augmentation solution. Automation is, after all, what many organizations have been dealing with since the industrial revolution. In AI automation, after a period of trial and error, machines take over a human task, whereas in AI augmentation, human capabilities are augmented

Artificial intelligence as a colleague  229

by intelligent machines that are capable of ‘learning’ based on previous interactions (as in the case of the Google search that learns from users’ previous searches). Further highlighting the complexity involved, our empirical research suggests that, in many cases, an AI solution labelled ‘automation’ may in fact also involve augmentation over lengthy time periods, as the solutions are implemented iteratively – and vice versa. This blurs the distinction between automation and augmentation further, and leads us to question the practical value of the typology in at least some cases. Often, not even the so-called ‘off the shelf’ or ‘plug and play’ AI-enabled automation solutions are ready for use, but require further development, given the difficulty in automating the more complex human tasks in a straightforward manner (Einola and Khoreva, 2023). The paradox perspective also entails the coexistence of humans and AI being an iterative or recursive process. Through both automation and augmentation, people and AI become so closely intertwined that they collectively exhibit entirely new, emergent behaviours that neither does individually (Amershi et al., 2014; Floridi and Taddeo, 2016; Beer, 2017). The coexistence of people and AI leads to the emergence of hybrid organizational settings, where it may be difficult or even impossible to distinguish between people and AI or their respective learnings and actions. In working through this complexity, it becomes apparent that people are arguably no longer the only agents in organizational settings. The use of AI for various tasks implies that AI is now not a simple artefact but a new class of organizational member (Floridi and Sanders, 2004). While AI has fundamental limitations, its actions enjoy far-reaching autonomy, because people delegate knowledge tasks to AI and allow it to act on their behalf (Rai et al., 2019). If we consider automation and augmentation interdependent, the interdependence inevitably extends towards people working with these AI solutions. The paradox perspective further implies that people and AI also interact on the same or closely related tasks. People shape AI through their daily choices, actions and interactions by defining objectives, setting constraints, generating and choosing training data and providing AI with feedback (Deng et al., 2017). Simultaneously, AI shapes people’s behaviour by informing, guiding and steering human judgment (Lindebaum et al., 2020; Moser et al., 2022). In our view, the value of the paradox perspective lies in the conceptual move from simple ‘either/or’ thinking (either humans or machines, either automation or augmentation) to a more encompassing ‘both/and’ mode (both humans and machines, both automation and augmentation). Accepting and embracing this duality allows us to recognize that the different AI solutions and humans are both complementary and contradictory, yet interdependent. This invites people and machines to coexist and mutually develop, as both people and AI learn in their own distinct way. Understanding how organizational members experience this coexistence is key, especially for organizations embarking on strategic, long-term AI projects involving multiple interlinked AI solutions and large numbers of staff in different roles.

230  Research handbook on human resource management and disruptive technologies

PATTERNS OF COEXISTENCE OF PEOPLE AND ARTIFICIAL INTELLIGENCE We have now introduced the term coexistence between people and AI, and employed the paradox perspective to demonstrate the complexity inherent to this coexistence. Next, we explore patterns of this coexistence in organizational settings. The Mundane and Incremental Nature of Artificial Intelligence Most of today’s AI solutions implemented in organizations are far from how AI was imagined, for instance, in sci-fi movies and futuristic novels (see Lindebaum et al., 2020). The organizational reality is much more mundane. Many organizations have been driving technology change projects since the advent of the personal computer. Many organizational members, rather than being swept away by an AI-propelled technology revolution targeting singularity (a possible future state where it is impossible to distinguish machine from human), are coping with slow technology evolution – and change fatigue. Rather than getting carried away with perhaps overly optimistic views of what AI can possibly achieve in the near or far future, we suggest that practitioners and researchers alike take a serious interest in actual organizational practice. Rather than fixating only on strategic or managerial aspects, they need to focus also on operations, look at what happens in current AI projects as they are rolled out, and consider how AI is being perceived by those organizational members who need to learn to coexist with the solutions in their daily work. Automation and Augmentation are Intertwined As discussed earlier, there is typically no automation ‘here’ and augmentation ‘there’, it is not ‘either/or’ (Einola and Khoreva, 2023). In contrast, both automation and augmentation have elements of both. Augmentation cannot be neatly separated from automation in an organizational setting deploying many incremental AI solutions, supporting the overall long-term organizational business goals. That is, augmentation is both the driver and the outcome of automation. Automation and augmentation are equally important puzzle pieces in constantly evolving company-specific technology roadmaps involving many AI solutions and other disruptive technologies. Furthermore, both AI solutions are inseparable parts of the same ‘picture’ needed to move organizations forward in their long-term AI implementation processes, including corresponding changes in jobs, skills and tasks conducted by both people and AI (Einola and Khoreva, 2023). Over-glorifying augmentation and downplaying automation, creating artificial walls between them and losing sight of the bigger picture, is a real risk for the harmonious coexistence of people and AI.

Artificial intelligence as a colleague  231

No Clear Cutovers between People and Artificial Intelligence There are seldom clear cutovers between people and AI in organizational settings. The interdependence between automation and augmentation extends towards people working with AI (Smith and Lewis, 2011; Schad et al., 2016; Raisch and Krakowski, 2021). Moreover, the coexistence between people and AI is not monolithic (Kaplan and Haenlein, 2019); rather, it is a constantly changing quest, nested in a given organizational setting and situated practice, comprising people designing AI solutions (i.e. technology developers and experts, vendors), those working with the solutions (i.e. frontline employees) and those in charge of the business (i.e. managers, business leaders). Even for what appears to be a simple AI-enabled automation solution to start functioning properly, there may be long periods of overlap where members in a focal organization need to collaborate tightly, not only amongst themselves, but also with technology partners, suppliers and even customers to help the solution work more effectively, and for advanced machine learning to become possible (Einola and Khoreva, 2023). This puzzle inevitably changes over time, and, consequently, the nature of the coexistence between people and AI transforms (Einola and Khoreva, 2023).

IMPACT OF COEXISTENCE ON JOB ROLES Much has been written about people losing jobs to AI and other disruptive technologies, and of AI doing in the future many more of the jobs presently done by humans (e.g. Leetaru, 2016). The counter-rhetoric is that AI does not have the volition or will to ‘take’ jobs from humans (Fleming, 2019), other than in fiction and Hollywood movies. The decision power and responsibility still lie solely with humans. The issue of some jobs disappearing to be replaced by others is hardly a new phenomenon, as job roles have been changing throughout history due to new technology or changes in processes, lifestyles and job design. Many of the jobs that were common in the late nineteenth century no longer exist, such as wagon-maker, telegraph operator or milkman, and new ones are constantly being created. It is clear thus far that AI is changing many jobs, and it is important to understand how AI impacts jobs and everyday organizational practice, rather than pander to speculation and sweeping suggestions of job losses. Understanding and documenting employee competences and planning for future skill gaps is a useful first step often neglected in the midst of all the change and strategy work. Human Job Roles are Gradually Changed But Not Necessarily Replaced Coexistence between people and AI causes important changes and re(creations) in human job roles (Brynjolfsson and McAfee, 2014, 2017; Faraj et al., 2018). Although domain expertise remains relevant to organizational members in educating and challenging AI, the coexistence leads to institutionalized knowledge – for instance, in the

232  Research handbook on human resource management and disruptive technologies

form of AI solutions and intellectual property – which is often in some ways superior or different to individual experts’ knowledge (Raisch and Krakowski, 2021). At the same time, general human skills that complement AI, such as creativity, common sense and advanced communication (Davenport and Kirby, 2016), as well as integration skills such as AI literacy, become more important (Daugherty and Wilson, 2018). Over time, as organizational members learn to use AI solutions and the technology is developed further, people also learn more about the limits and possibilities of AI, triggering further changes and re(creations) in job roles. However, these changes in job roles are not necessarily linear and straightforward in direction (Einola and Khoreva, 2023). Even though AI is occasionally portrayed as a new agent and a rival to people in the emerging surveillance capitalist system (Zuboff, 2019), it does not necessarily capture tasks previously done by people, at least not in the foreseeable future. In fact, many seemingly mundane tasks involve creativity and situation-sensitive, trust-based human interaction, as well as tacit knowledge embedded in employee routines. In conducting an empirical study of one of the largest media agencies in Finland, we saw concerning signs of narrowing down and verticalization of job roles focusing more on the AI itself, as people learned skills such as coding and familiarity with Google proprietary tools, and much less on the business and its human side (Einola and Khoreva, 2023). This was particularly visible in a dichotomy arising between the employees working in the digital and traditional media groups. Digital media employees were focused on sharpening their technical skills, thus engaging in AI-induced up-skilling. At the same time, they lacked knowledge of and often even interest in the broader media landscape and business issues, skills that the management had identified as a must to win future customer contracts. This contrasted with employees working on the traditional media. While at times admitting to a lack of advanced digital skills, they had a good oversight of broader business and practical matters, had mastered many types of traditional media channels and had close ties with people working in partner and customer companies. All these qualities were needed to do their jobs successfully and proved to be extremely difficult to capture in AI solutions. Establishing ‘clear rules’ translatable into algorithms may not be that simple (Raisch and Krakowski, 2021) and the value of human tacit knowledge should not be underestimated. The Re-Humanizing Perspective Needs to be Revisited AI does not necessarily re-humanize the work of people, as partially prescribed by the re-humanizing perspective (Armstrong, 1973). The coexistence of people and AI does not automatically shift the work of people from repetitive and monotonous tasks, unreflectively labelled as unwanted, towards those that are more creative and fulfilling (Daugherty and Wilson, 2018) – nor is this necessarily a goal to be pursued. Returning to our media agency case study, we observed that, while top managers were eager to automate all possible tasks, enabling personnel to be creative and more productive, employees were often pronouncedly sceptical and weary of constantly

Artificial intelligence as a colleague  233

being inspired and coming up with novel ideas (Einola and Khoreva, 2023). The so-called monotonous tasks allowed employees to get some much needed distance and rest from the pressure to be innovative and creative. This rest in the form of monotonous tasks fuelled their creative capabilities. As one media planner put it: ‘no one can create, create and create all the time!’ Similarly, shortening the planning, campaign execution and reporting cycles was experienced as a double-edged sword. Increased efficiency led to a shortage of time to think and reflect, and to the workplace becoming a more hectic, chaotic and stressful habitat. Different Job Role Changes for Different Organizational Groups The coexistence of people and AI causes concrete changes in the job roles of different organizational groups. While everyone will need to learn to use AI solutions, our extensive fieldwork suggests that most people in many organizational settings are already accustomed to different tools, and learning to use AI may not be that different (Einola and Khoreva, 2023). As the coexistence of people and AI becomes an anchored organizational reality that expands and deepens over time, the changes triggered by AI may not be dramatic and immediate but incremental. These incremental changes also generate flowing learning. The meaning and nature of these changes, however, varies among different organizational members. For executives and business leaders, human/AI coexistence may be essential to the long-term prosperity and viability of their company, hence something that other organizational members willing to remain in the company need to learn to accommodate. Engaged in large-scale AI implementation projects, much of top managers’ and business leaders’ attention may shift from operations and managerial work to the development and implementation of organizational technology strategies and roadmaps, as well as the creation and nurturing of the nascent overall working ecosystem with technology partners, customers and others. The risk here is increased oversight of operations and the decoupling of managerial from employee reality, creating tension in the organization (Einola and Khoreva, 2023). Managers, in turn, may gain more managerial responsibilities needed to drive the change, such as communication, coordination, interpretation and ‘translation’ of AI solutions to different target groups, such as employees, customers and business partners, to ensure everyone is on board (Einola and Khoreva, 2023). These extra tasks may shift their attention from running the change projects on the ground, leading to further work in terms of troubleshooting and firefighting, if things do not go as planned (perhaps due to deficient planning). As for the employees, working hand-in-hand with AI requires much time and effort spent refining and aligning processes to manifest the possibilities offered by AI solutions. While executives and managers may assume some tasks to be quite straightforward, simple and easy to model and automate, the picture may look different at the employee’s desk. Instead of handing over tasks to AI, the employees’ focus shifts towards testing, fixing and working with AI (Einola and Khoreva, 2023). Thereby, an initial phase with employees and AI solutions executing overlapping

234  Research handbook on human resource management and disruptive technologies

tasks may take longer than expected; raising the question of in whose time will all this testing and improvement work be done – employees’ working hours or their own personal time? The danger here is that employees are left alone to resolve this dilemma, as they juggle multiple, sometimes conflicting or competing demands and commitments, leading to dissatisfaction and conflict. Employees’ tacit knowledge of colleagues, customers and the business is frequently required to resolve deviations, making, for example, an AI-enabled automation solution look like a mixed automation-augmentation application regularly requiring employee intervention. For example, in our case study of a media agency implementing several AI solutions simultaneously, the automation solution needed to be refined by technology partners based on user feedback (Einola and Khoreva, 2023); then human users needed to learn to work with the newly fixed automation solution. This required a lot of employee time and effort that had been downplayed by the management assuming the employees’ job was simpler and lower-skilled than it actually was.

CONCLUSION AND RECOMMENDATIONS FOR ACTION The abovementioned insights and considerations lead to several crucial recommendations for action. First, business leaders need to give balanced consideration to the entire technology roadmap and all the AI projects in which they engage (Einola and Khoreva, 2023). A ‘simple’ AI-enabled automation solution may involve much more iteration in the form of integrated automation and augmentation, as well as employees’ time and effort, than initially thought. Both automation and augmentation should be in focus, and have the same status on the management and employees’ agenda. Otherwise, there is a danger of over-glorifying technologically advanced front-office AI-enabled augmentation solutions, and not seeing much interest in the practicalities of what it takes to run long, but equally important, automation-related back-office projects. Second, business leaders and HR professionals need to remember that AI implementation may take much more time and effort than first expected (Einola and Khoreva, 2023). People and AI need to change slowly, in tandem, as learning occurs. Neither automation nor augmentation can be delegated to the IT department or technology partners and be expected to roll out without all members involved being on board. As a direct consequence, both management and HR departments need to plan AI implementations thoroughly, and allocate sufficient resources, not least when it comes to employee time. Moreover, human-to-human interaction around AI must be both a managerial and employee priority. Third, business leaders and HR professionals need to carefully map and plan for job role changes and competence shifts, and ask questions such as: What are our people really doing? What drives their motivation (or lack thereof)? What skills are needed? How do we manage the competence shift? What tacit knowledge is involved? Have we allocated enough employee time to make an AI implementation

Artificial intelligence as a colleague  235

possible? And perhaps most importantly: How do we ensure our people can peacefully coexist with their new artificially intelligent colleagues? Fourth, scholars need to ground their studies in empirics and organizational practice. This requires in-depth empirical studies, long-term field engagement and the use of multiple methods. Business leaders, perhaps somewhat differently to technology experts and unconditional AI enthusiasts, do not first choose between automation and augmentation (Raisch and Krakowski, 2021) as a first-hand solution to an overarching problem, but seek to address real-life organizational challenges with the help of available AI solutions or by other means. Consultants, authors of best-selling books (see, for instance, Brynjolfsson and McAfee, 2014, 2017; Davenport and Kirby, 2016; Daugherty and Wilson, 2018) and scholars alike rush companies to change their culture to succeed in adopting a data-centric view, and even to abandon old practices, such as reliance on human experience and intuition. For many employees, the future does not look bright, and there will be winners and losers; those who are able to ride the digital wave and those who fall. Many consider essential human qualities a hurdle to overcome through big data and machine intelligence. However, this revolution is not going to succeed or lead to better organizations without thinking through the role people have in turning the ongoing change into the transformational force it can become. This chapter has depicted a richer, more nuanced and pragmatic version of reality, where employees are challenged to learn to coexist peacefully with AI as a new type of a colleague. We have introduced the idea of AI emerging as a new type of organizational member to be understood and managed differently than earlier technologies such as fax machines, cars, personal computers or Microsoft Office applications. The key difference here is that while people use these earlier technologies as tools, people and AI need to engage with each other much more closely. Many functions of AI are not even visible or comprehensible to people who interact with it. This engagement leads to changes and re(creations) of human job roles. Consequently, the boundary between people and AI blurs. Thus, we suggest that both scholars and practitioners should consider this new organizational phenomenon a study of coexistence of people and AI. Coexistence is particularly suited to this context, because it is an organic term, neither jumping on the AI-hype bandwagon nor considering it as a threat to humanity. Due to inherent complexity in the study of human/AI coexistence, we believe it is particularly promising to extend the paradox theory to include not only different types of AI solution but also people who interact with these solutions.

REFERENCES Amershi, S., Cakmak, M., Knox, W. B., and Kulesza, T. (2014). Power to the people: the role of humans in interactive machine learning. AI Magazine, 35(4): 105–20. Armstrong, R. L. (1973). The rehumanization of work. Social Theory and Practice, 2(4): 459–73.

236  Research handbook on human resource management and disruptive technologies

Bar-Tal, D. and Bennink, G. (2004). The nature of reconciliation as an outcome and as a process. In Y. Bar-Siman-Tov (ed.), From Conflict Resolution to Reconciliation (pp. 11–38). New York: Oxford University Press. Beer, D. (2017). The social power of algorithms. Information, Communication & Society, 20: 1–13. Brynjolfsson, E. and McAfee, A. (2014). The Second Machine Age: Work, Progress, and Prosperity in a Time of Brilliant Technologies. New York: W.W. Norton. Brynjolfsson, E. and McAfee, A. (2017). The business of artificial intelligence. Harvard Business Review, July. Daugherty, P. and Wilson, H. J. (2018). Human + Machine: Reimagining Work in the Age of AI. Boston, MA: Harvard Business Review Press. Davenport, T. H. and Kirby, J. (2016). Only Humans Need Apply: Winners and Losers in the Age of Smart Machines. New York: HarperCollins. Deng, Y., Bao, F., Kong, Y., Ren, Z., and Dai, Q. (2017). Deep direct reinforcement learning for financial signal representation and trading. IEEE Transactions on Neural Networks and Learning Systems, 28: 653–64. Einola, K., and Khoreva, V. (2023). Best friend or broken tool? Exploring the co-existence of humans and artificial intelligence in the workplace ecosystem. Human Resource Management, 62(1): 117–35. Faraj, S., Pachidi, S., and Sayegh, K. (2018). Working and organizing in the age of the learning algorithm. Information and Organization, 28(1): 62–70. Fleming, P. (2019). Robots and organization studies: why robots might not want to steal your job. Organization Studies, 40(1): 23–38. Floridi, L. and Sanders, J. W. (2004). On the morality of artificial agents. Minds and Machines, 14: 349–79. Floridi, L. and Taddeo, M. (2016). What is data ethics? Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences, 374: 1–4. Glikson, E. and Woolley, A. (2020). Human trust in artificial intelligence: review of empirical research. Academy of Management Annals, 4(2): 627–60. Haenlein, M. and Kaplan, A. (2019). A brief history of artificial intelligence: on the past, present, and future of artificial intelligence. California Management Review, 61(4): 5–14. Huang, M.-H., Rust, R., and Maksimovic, V. (2019). The feeling economy: managing in the next generation of artificial intelligence. California Management Review, 61(4): 43–65. Kaplan, A. and Haenlein, M. (2019). Siri, Siri, in my hand: who’s the fairest in the land? On the interpretations, illustrations, and implications of artificial intelligence. Business Horizons, 62(1): 15–25. Kriesberg, L. (1998). Coexistence and the reconciliation of communal conflicts. In E. Weiner (ed.), The Handbook of Interethnic Existence (pp. 182–98). New York: Continuum. Lanier, J. (2014). Who Owns the Future? New York: Simon and Schuster. Latour, B. (1992). Where are the missing masses? The sociology of a few mundane artifacts. Shaping Technology/Building Society: Studies in Sociotechnical Change, 1, 225–58. Leetaru, K. (2016). Is Elon Musk right and will AI replace most human jobs? Forbes, November 8. Available at https://​www​.forbes​.com/​sites/​kalevleetaru/​2016/​11/​08/​is​-elon​ -musk​-right​-and​-will​-ai​-replace​-most​-human​-jobs/​#6c6b41a860f4. Lindebaum, D., Vesa, M., and den Hond, F. (2020). Insights from the machine stops to better understand rational assumptions in algorithmic decision-making and its implications for organizations. Academy of Management Review, 45(1): 247–263. Mori (1970). The Uncanny Valley. Energy, 7: 33–5. Moser, C., den Hond, F., and Lindebaum, D. (2022). Morality in the age of artificially intelligent algorithms. Academy of Management Learning & Education, https://​doi​.org/​10​.5465/​ amle​.2020​.0287.

Artificial intelligence as a colleague  237

Nilsson, N. J. (1971). Problem-Solving Methods in Artificial Intelligence. New York: McGraw-Hill. Nilsson, N. J. (2009). The Quest for Artificial Intelligence. Cambridge: Cambridge University Press. Rai, A., Constantinides, P., and Sarker, S. (2019). Next-generation digital platforms: toward human-AI hybrids. MIS Quarterly, 43: 3–9. Raisch, S. and Krakowski, S. (2021). Artificial intelligence and management: the automation-augmentation paradox. Academy of Management Review, 46(1): 192–210. Rothstein, R. L. (ed.). (1999). After the Peace: Resistance and Reconciliation. Boulder, CO: Lynne Rienner Publishers. Sayes, E. (2014). Actor–network theory and methodology: just what does it mean to say that nonhumans have agency? Social Studies of Science, 44(1): 134–49. Schad, J., Lewis, M. W., Raisch, S., and Smith, W. K. (2016). Paradox research in management science: looking back to move forward. Academy of Management Annals, 10: 5–64. Smith, W. K. and Lewis, M. W. (2011). Toward a theory of paradox: a dynamic equilibrium model of organizing. Academy of Management Review, 36: 381–403. Snell, S. and Morris, S. (2021). Time for realignment: the HR ecosystem. Academy of Management Perspectives, 35(2). Weiner, E. (1998). Coexistence work: a new profession. In E. Weiner (ed.), The Handbook of Interethnic Existence (pp. 13–24). New York: Continuum. Whittaker, D. J. (1999). Conflict and Reconciliation in the Contemporary World. London: Routledge. Zuboff, S. (1988). In the Age of the Smart Machine: The Future of Work and Power. New York: Basic Books, Inc. Zuboff, S. (2019). The Age of Surveillance Capitalism: The Fight for a Human Future at the New Frontier of Power. New York: Public Affairs.

17. Platform work inside organisations: an exploration of tensions in intra-organisational labour platforms Philip Rogiers, Jeroen Meijerink and Stijn Viaene

INTRODUCTION This chapter presents an empirical study of the organising tensions in and around intra-organisational labour platforms (IOLPs), which are digital platforms that match employees to part-time, fixed-term work opportunities inside the bounds of organisations (Rogiers et al., 2020, 2021). By affording employees to move ‘fluidly’ across internal boundaries and work units (Altman et al., 2021), IOLPs disrupt the structure and functioning of contemporary organisations. They blur internal structures and human resource management (HRM) practices, interrupt primary work processes and open new ways for internal mobility and career advancement. While IOLPs have been identified as ‘the central organising principle for making more people more valuable in more organisations’ (Kiron et al., 2020), their disruptive potential also causes problems. For instance, incongruences can arise between how employees, in their different platform roles, make sense of and act on the IOLP’s action possibilities. Such incongruences can incarnate in organising tensions (i.e. the experience of competing demands, the tug-of-war between alternative options; Smith and Lewis, 2011) within and surrounding the IOLP that, if left unmanaged, risk undermining the value and effectiveness of this new labour concept (Rogiers et al., 2020, 2021). In this chapter, we probe into IOLP tensions and ask: Which tensions emerge within and around IOLPs? Why do they arise? And how can IOLP designers address such tensions? Our empirical study of IOLPs builds understanding of the disruptive potential of this new platform type (through our focus on organising tensions) and the consequences for organisations and HRM practices. We find that these platforms face unique tensions and deploy novel mechanisms to keep these in check. Tensions relate to conflicting goals as participants enact different roles in the intra-organisational platform as well as within the wider organisation that employs them, but nevertheless are interdependent to make the platform succeed. Our study intends to be a stepping stone for a new stream in the digital labour platform and HRM literature that considers the managerial and organisational implications of platform labour in organisations. This chapter is structured as follows. We briefly review the literature on IOLPs and explain how they work, after which we explain our choice for an affordances and 238

Platform work inside organisations  239

constraints lens and provide a conceptual background to this theory. After outlining our study context and methods, we apply the affordances and constraints lens to study IOLP tensions in the case context of STRETCH, an IOLP in a large/multi-department governmental organization. Given the platform’s growth and adaptation since its inception in 2013 (including more than 1000 completed projects and 8000 registered employee profiles), STRETCH provides a notable context to investigate how IOLPs operate as a new work model and manage emerging tensions. We conclude by highlighting our study’s theoretical implications and recommendations for practical action.

LITERATURE OVERVIEW AND CONCEPTUAL FRAMEWORK Intra-Organisational Labour Platforms: A New Organisational Phenomenon The last decade has seen a stellar rise in platform labour, a new employment form characterised by often temporary and on-demand work engagements (Spreitzer et al., 2017; Stanford, 2017; Ashford et al., 2018) facilitated by a digital labour platform intermediary (Kuhn et al., 2021; Meijerink et al., 2021). Examples are Uber, Upwork, or Fiverr that have quickly stepped in to replace the standard employment relationship (Meijerink and Keegan, 2019; Duggan et al., 2020; Vallas and Schor, 2020). Similar to traditional organisations, these platforms keep control over the allocation of tasks and the collection of data and revenues; but they differ by ceding control over the specification of work methods, control over work schedules and the labour of performance evaluation (Vallas and Schor, 2020). While the literature has mainly focused on labour platforms in the ‘gig economy’ (Maffie, 2020), organisations, too, have noted the success of these platforms, and some of them have experimented with analogous approaches for internally organising their permanent workforce (Kiron et al., 2020) through the operationalisation of IOLPs. IOLPs bring a new organisational artefact to the fore, a digital labour platform that facilitates the matching of permanent employees to work projects across internal boundaries in the organisation (Rogiers et al., 2020, 2021; Schrage et al., 2020). The matching occurs in a digital platform environment that is designed and maintained by IOLP designers. This affords IOLP participants to operate a smartphone application or website that enables them to post and react to project assignments. Moreover, IOLP designers provide project posters and project workers, who respectively represent the labour demand and supply side of IOLPs, with the digital means to connect to peers, work remotely on projects and store shared documents (Rogiers et al., 2021; Schrage et al., 2020). By allowing employees to post, react to and organise part-time, fixed-term work projects, IOLPs effectively assume responsibility for brokering the supply and demand of deconstructed work elements (i.e. tasks, projects) within the bounds of the organisation (Boudreau et al., 2015; Gantcheva et al., 2020; Schrage et al., 2020; Altman et al., 2021; Boudreau, 2021). Although project posters who

240  Research handbook on human resource management and disruptive technologies

propose projects via an IOLP retain coordination capability over their projects, they now share this role with the IOLP, with the latter orchestrating the matching process through which intra-organisational projects are resourced and organised (Weller et al., 2019). This makes the IOLP a powerful artefact that can disrupt the organisational status quo, but also create new tensions in organisations. To understand why tensions arise and whether and how IOLPs can address these, we adopt the lens of technology affordances and constraints theory. Technology Affordances and Constraints: Background to Our Conceptual Lens To understand the disruptive nature of IOLPs, we draw on the theory of technology affordances and constraints (Majchrzak and Markus, 2012). Technology affordances and constraints theory offers a fitting perspective to understand the disruptive potential of OLPs, as it can explain how such potential is enacted through human-technology interactions in the modern workplace (Majchrzak and Markus, 2012). In line with the tenets of affordances and constraints theory, we assert that disruption both enables something new (i.e. a technology affordance: the possibilities of action offered by a technological artefact to organisational actors; Leonardi, 2013; Leonardi and Vaast, 2017), but also hinders existing actions and routines through technology constraints (i.e. how a technological artefact hinders organisational actors from accomplishing a particular goal; Volkoff and Strong, 2013). The concepts of technology affordance and constraint stem from the more general notions of affordance and constraint introduced in 1966 by James J. Gibson, who conceptualised affordances as the possibilities of action that the environment’s materiality offers to an actor – or an animal in his original writings, which focused on actor-artefact interactions in the biological world (Gibson, 1979). Our choice for an affordances lens reflects our view of IOLPs as multisided platform environments (Hagiu and Wright, 2015) wherein multiple actors pursue different and possibly conflicting goals (Volkoff and Strong, 2013; Leonardi and Vaast, 2017). However, actors may enact the IOLP’s affordances in incongruent ways, resulting in organising tensions (i.e. the experience of competing demands, the tug-of-war between alternative options; Smith and Lewis, 2011). Tensions may arise as people come from different organisational units and hold different backgrounds, commitments and responsibilities and motives for engaging in the IOLP. Besides the possibility of organising tensions arising from participants’ incongruently enacted affordances, tensions can also arise between the affordances shaped by IOLP designers and the way they are enacted by platform participants. Designers’ infusing of intentions into the material and social dimensions (Faraj and Azad, 2012) of IOLPs shapes intended affordances through which they hope to proactively anticipate and shape participant behaviour. Participants may diverge from these intended affordances as they perceive them as individually constraining (Leonardi and Vaast, 2017) or as they are constrained by forces in their broader organisation (Volkoff and Strong, 2013). We label these incongruences as endogenous and exogenous incon-

Platform work inside organisations  241

gruences, respectively, thereby providing insight into how eventual tensions arise in IOLPs. It is further important to highlight that our technology affordances and constraints perspective differs from related approaches that analyse both material and social dimensions of technology, such as sociomateriality (Orlikowski, 2010; Leonardi, 2013; Cecez-Kecmanovic et al., 2014). While the latter interprets platforms as sociomaterial contexts that are inseparable from the actors that engage in them (Leonardi, 2013), technology affordances and constraints theory sees the material and social as connected but standalone dimensions (Faraj and Azad, 2012; Leonardi and Vaast, 2017). This distinction is critical, as it allows us to untangle the tensions that an IOLP brings forth, which requires distinct interventions in either the social and material dimensions of IOLPs. The social dimension of an IOLP relates to the collective of platform participants (both in their role as project poster and/or project worker) that affords and constraints the IOLP in realising its goals of matchmaker, while its materiality reflects the technological infrastructure (i.e. the digital platform) that affords and constraints the IOLP’s participants to enact their role as a project poster or project worker. It is the affordances and constraints that are at the root of tensions in IOLPs and that we explore in this study. Our study also shows how tensions can be addressed through distinct interventions in an IOLP’s social and material dimensions. Our affordance lens’ ability to disentangle these organising tensions and interventions helps us understand the significance of IOLPs, not just as a novel work context, but as a key artefact that carries and manages new tensions in the contemporary workplace.

METHOD Case Description This research builds on a case study of STRETCH (fictitious name, which we choose since the IOLP under study offered so called‘stretch assignments’ to employees), an IOLP inside a large/multi-department governmental organisation. STRETCH allows government employees to explore project-based work opportunities within the large/ multi-department governmental organization rather than full-time, in-role tasks for which workers bear formal responsibility and accountability. Formally, employees could take on these projects for a maximum of 20 per cent of their weekly work time without having to pause their full-time jobs. Taking a ‘one-government’ (interviewee #1) approach to the design of STRETCH was a deliberate choice of the platform designers: targeting a broad participant base across the multi-department governmental organization would ensure that enough posters and workers transacted on the platform. To achieve this aim, efforts were made to promote the platform among participant types, such as through presentations, workshops and weekly newsletters. However, a broader participant base also increased the chances for organising tensions later, which was a calculated risk that

242  Research handbook on human resource management and disruptive technologies

needed to be managed going forward. STRETCH’s 2013 pilot attracted a broad participant base as well as employees’ early enthusiasm; with individuals from over 50 government agencies signing up for over 100 projects in a variety of areas. While STRETCH initially aimed to offer only digital-themed projects, its designers quickly removed this restriction to draw even more post(er)s to the platform. As of January 2020, 730 unique projects had been posted (not including projects that have been reposted to recruit more workers), employees had submitted nearly 3000 applications and 1456 individuals had taken part in STRETCH projects. Data Collection We collected primary case data as part of a wider study of STRETCH between May 2018 and January 2021. In total, we conducted 59 semistructured online interviews at regular intervals during this observational period, including interviews with platform participants (i.e. project posters and project workers), platform designers and community managers, resulting in approximately 41 hours of audio recordings. Interviews were transcribed verbatim and checked for accuracy (Miles and Huberman, 1994). To triangulate the interview findings (Stake, 2000), we collected complementary data, including platform data, newsletters, data on GitHub, articles and blog posts on social media, information from informants’ LinkedIn profiles, government web pages, videos, blog posts on government media, strategic documentation, internal working documents and presentations. Data Analysis We inductively analysed our interview transcripts and complementary data (Glaser and Strauss, 2017). Aided by the qualitative data analysis software MAXQDA, we started by iteratively parsing out open and in vivo codes (Strauss and Corbin, 1997; Charmaz, 2006). Starting without preconceived notions about what the codes should be, our first-order codes revolved around STRETCH’s technological design and participants’ experiences with the platform. Through repeated reading and analysis, we integrated and grouped first-order codes into more abstract categories (Strauss and Corbin, 1997; Gioia et al., 2013), informed by our theory lens. We identified both affordances and constraints categories in our data. We further noticed that interviewees addressed the platform’s material and social dimensions as they shared their experiences with STRETCH; a distinction that we included in our coding process. As we coded our data through this lens, we eventually noticed how the affordances enacted by one platform actor could be incongruent with the affordances of other actors. Said differently, we discovered that the affordances enacted by one type of participant could be constraints for other types of participants. Accordingly, we established links between our different affordance and constraint codes, from which our aggregate construct of organising tensions emerged.

Platform work inside organisations  243

Table 17.1

Organising tensions in IOLPs

Organising tension

Tension description

Tension characteristics (endogenous/

Tension one:

The experience of competing demands between Endogenous, arising from incongruently

exogenous) developmental affordances project workers’ goal to seek development and

enacted affordances of participants,

vs. low-cost labour

project posters’ goal to seek low-cost labour

where the affordances enacted by one

sourcing affordances

sourcing in the IOLP

Tension two: marketplace The experience of competing demands

type of participant constrain others Endogenous, arising from incongruences

affordances vs. community between the goal of IOLP designers to build

between designers’ intended affordances

affordances

a transactional market of temporary projects

and participants’ enacted affordances,

and participants’ goal to connect with and feel

where participants experience designers’

related to others beyond temporary projects

intended affordances as constraining and as such, diverge from these intentions in their enacted affordances

Tension three: new role

The experience of competing demands between Exogenous, arising from incongruences

affordances vs. traditional the goal of IOLP designers to enable role role affordances

between intended and enacted

experimentation and participants’ tendency to

affordances, where participants’

enact traditional role patterns in the IOLP

enactment of designers’ intentions is constrained by forces originating outside the platform environment

FINDINGS Our data revealed three organising tensions in IOLPs, depicted in Table 17.1. The first tension shows how competing demands can arise from incongruently enacted affordances, where the affordances enacted by one type of platform participant can be a constraint for others. The second tension further shows how tensions can originate from endogenous incongruences between intended and enacted affordances, where participants experience designers’ intended affordances as constraining and, as such, diverge from these intentions in their enacted affordances. Finally, the third tension exhibits how tensions can emerge from exogenous incongruences between intended and enacted affordances, where participants’ enactment of designers’ intentions is constrained by forces originating outside the platform environment. The following sections demonstrate why and how each tension comes about, as well as the possibilities and limitations for IOLPs designers to defuse these tensions. Tension One: Developmental Affordances vs. Low-Cost Labour Sourcing Affordances Because of the STRETCH designers’ initial choices to target a broad participant base, the platform enjoyed a steady source of project posters and workers from the start. Nonetheless, our data show how tensions can flare up in an IOLP, ‘caus[ing] conflict’ (#10) between workers and posters, and challenging designers’ intention to keep participants committed to the platform. Referring to these organising tensions,

244  Research handbook on human resource management and disruptive technologies

a designer expressed: ‘The value of a [participant] tends to go down over time. They just feel like, “Eh, I haven’t had that great of an experience”’ (#1). At the root of such tensions lay the goals of workers and posters, which differed. On the one hand, workers mostly came to STRETCH to experiment with new work through fixed-term projects. A worker expressed: ‘I’m getting a new adventure. I get to learn new things and grow a lot more’ (#31). IOLPs thus enabled workers to step outside of their daily routine and seek new developmental opportunities, which made up workers’ primary goals. Project posters were often more outcome-oriented and were primarily looking for workers that already possess the required expertise. They commonly viewed the IOLP as an opportunity to shop for an extra pair of hands for a specific task for which they had clear outcome expectations. For instance, a poster stated: Typically, projects fall in the realm of activities where maybe you don’t have [the] in-house expertise or where it’s something that’s nice-to-have, not need-to-have … So, to outsource it to someone who has the needed expertise is really valuable. And certainly, when you have expertise gaps, it’s extraordinarily helpful because I see it as a way to advance work through resource leveraging. (#23)

While the goals of project workers and posters were not necessarily contradictory, either workers or posters could enact affordances in incongruent ways, where the affordance enacted by one side constrains the other – creating tension in the IOLP. For example, as posters enacted the IOLP’s affordance to outsource tasks, some posters did so in a way that constrained workers’ goals to develop themselves through social engagement and receive guidance from posters. A worker emphasised how she experienced this constraint by stating, ‘I don’t have good mentorship or good peers to bounce ideas off of’ (#13). Another worker elaborated how STRETCH, for her, was about ‘an opportunity to meet someone else or [receive] a little bit of coaching’ (#3) but how she felt constrained by project posters who ‘thought that it was very cumbersome. And they didn’t have the time for that.’ Similarly, workers’ enactment of the IOLP’s affordance to experiment with new work at times also constrained posters’ goal to source a reliable extra pair of hands through the platform. For instance, some posters complained about workers not following through on projects because of waning interest or motivation: We picked up someone … and he needed to do some hours with another entity. And so he used [STRETCH] working with us. And [our] expectation was there was going to be an end-product. And because he was just working the hours he needed, he passed off some work product, but [it] wasn’t the final product. (#19)

Other posters concurred, ‘They promised a lot and [did] not deliver!’ (#7). Notwithstanding these examples, we also came across instances where workers’ and posters’ enactment of affordances did not create tension. Yet, we found that these cases typically applied to participants who were aware of and compromising towards

Platform work inside organisations  245

the other side’s goals in the platform. For instance, one such participant – a project poster – explained: You have to be motivated by the learning and development parts of it to make [STRETCH] really worth your time … I’m willing to put in the extra time to do it and know that the product at the end might take a lot of polishing from me to get it to the place that I want it to be. But it also allows me to do special projects that I maybe wouldn’t have been able to do otherwise. But [posting projects] requires a lot of upfront thinking, and you have to be very intentional about it. (#38)

As not all STRETCH’s participants exhibited this understanding toward others, designer interventions were needed to address these emergent tensions between workers and posters in the IOLP. These interventions focused on either the platform’s material or social dimensions; two distinct but related dimensions captured by our affordance lens (Volkoff and Strong, 2013; Leonardi and Vaast, 2017). To address the tensions arising from incongruently enacted affordances, the IOLP designers first introduced changes in the platform’s material dimension. One example was the introduction of a new performance evaluation feature to manage interactions between workers and posters. Where freelance platforms commonly use worker ratings to sanction poor performance (Kuhn and Maleki, 2017), STRETCH introduced ‘exceptional performance badges’ that afforded project posters a view of workers’ prior performance while safeguarding workers’ affordance to keep learning and experimenting without being punished. The IOLP designers also directly intervened in the platform’s social dimension, for example, through workshops and awareness sessions, where designers invested in training posters on how to identify and plan projects. The project posting process also included questions and tips that prompted project posters to identify and plan projects in a worker-centric manner, for example, by underscoring that ‘Every [project] is a learning opportunity’. Complementary interventions in the platform’s social dimension included targeted communication and perception management efforts, such as STRETCH’s sharing of success stories that signalled the platform’s developmental nature to posters and workers. A designer elaborated on these efforts: There’s all kinds of success stories … people that have developed their careers and got out of what they were doing and tried some new things and were able to change their jobs based on it … we’re do[ing] a number of little testimonial videos once people have completed opportunities, the value that they got. Tell their stories. We’re writ[ing] little articles about that. (#1)

Finally, to further enhance the platform’s ability to intervene in its social dimension, the role of community manager was introduced in STRETCH. These individuals served as local liaisons in each participating department and facilitated participant activity and adoption of norms on the platform. A community manager explained that her role involved signalling the legitimacy and developmental nature of the platform to participants:

246  Research handbook on human resource management and disruptive technologies

[I do] not just say, ‘Hey, go to this website.’ But [I] say, ‘Hey, go to this website, look at it. What we do, we’re a volunteering group.’ I explain to them it’s all about professional development, it’s supported, we’re allowed to use it. (#7)

Elaborating on her role in educating participants about ‘good platform behaviours’, the same community manager explained how she ‘show[ed] them the basic courtesies’ by making it clear to participants that they should not ‘do the things that the rest of the government does, which is ask[ing] for things and then never respond or go into a black hole’ (#7). Tension Two: Marketplace Affordances vs. Community Affordances STRETCH was intended to be a multisided platform (Hagiu and Wright, 2015) within a large/multi-department governmental organization, affording direct interactions between employees in a ‘marketplace of tasks’, so people could work ‘across the country, across agencies’ (#6). While empowering people was at the heart of STRETCH’s ‘one-government’ intentions, we found that the market-logic in designers’ intended affordances created tension with the goals of participants – especially project workers. The latter commonly stated how they felt constrained by the platform’s transactional affordances – or the possibilities for providers and requesters of labour to interact; a feeling which a participant voiced as: ‘That’s just labour … a job or tasking transactional thing’ (#27). Referring to these intended affordances, other participants cited the lack of community feel, such as interviewee #17, who stated: ‘I really don’t feel that the [STRETCH] platform is a community.’ Explaining her desire to connect with and feel related to others beyond temporary projects, the same interviewee continued by expressing the need for the platform to afford her ‘a framework or a structure for that particular area of knowledge to allow colleagues to share their expertise in a more ongoing way and build on each other’s expertise in a more ongoing way’. She stressed: ‘But that is not a project or a transaction. It’s more of this community building’ (#17). To overcome the limitations of designers’ intended affordances, several participants themselves acted on the possibilities of the platform’s social dimension to realise their goals of feeling and being connected to others. These individual efforts included establishing professional ties and friendships that outlasted the lifespan of projects, and the establishment of communities of practice beyond the platform, such as the ‘user experience design community’, the ‘data-savvy community’, the ‘cyber community’ and the ‘acquisitions community’. Participants’ engagement in such communities limited their coordination costs, allowing them to share similar organisational backgrounds, draw on similar values and ideas and speak a shared language (Nahapiet and Ghoshal, 1998). However, it also created endogenous incongruences with designers’ intended affordances in the IOLP, as these communities risked becoming increasingly detached from the IOLP. To smooth the endogenous incongruence that arose from participants’ purposeful deviation from designers’ intended affordances, the designers ended up intervening in both the material and

Platform work inside organisations  247

social dimensions of the platform. In the material dimension, new technology features were introduced to tie emerging communities of practice to the IOLP. These features included digital community portals where community members could post projects and provide information and updates about their activities. Examples were the portals created for the cyber security community, for the digital user experience design community, for the ‘data-savvy’ community, and for the ‘acquisitions’ or for the contracting community. A designer explained: We took to heart that there was a need within a discipline, within a type of career, like the acquisitions and the cyber [communities]. They wanted to do things that really helped develop their cadre of workers … [STRETCH] launched an acquisition portal or section for acquisition opportunities … the acquisition community, which is cross-agency … is hoping that by providing developmental opportunities that then will show on your profile … that you have a way to see and for others to see that you’re confident or proficient in a skill. (#1)

Designers also intervened in the platform’s social dimension to signal that STRETCH was more than a transactional labour platform. Efforts were made to signal to participants that they were key enablers of STRETCH’s ‘one-government’ mission as they contributed to the platform’s original value propositions of ‘solving needs across government’ (#1) and ‘better serv[ing] the citizens of [our country]’ (#20). A designer elaborated: What we attached to [the platform] was this mission-oriented[ness]. That people really believed in the mission of the government or their agency, and that was an incredible positive characteristic that led this culture change … ‘make a difference’, ‘join a [country-wide] network, solving needs across government’. (#5)

This mission-statement was highlighted on the STRETCH homepage and in weekly newsletters, information sessions and presentations. We found how this communal aspect resonated with newer participants, such as interviewee #26, who said: ‘It virtually made me feel like I was part of a larger group, even though I don’t really know 90% of these people. It felt like, “Hey, we’re all part of this movement”, right?’ Tension Three: New Role Affordances vs. Traditional Role Affordances While experimentation with new work activities and role patterns often takes place outside one’s primary organisation (e.g. through side-gigs, hobby projects, or dual careers; Campion et al., 2020; Sessions et al., 2021), the IOLP designers constructed STRETCH as a playground inside the organisation that affords experimentation with alternative roles and identities. Yet we found that these intended affordances were incongruent with the constraints exerted by organisational norms and routines (Nelson and Winter, 1982; Grant, 1996; Jones, 1986) external to the platform, but to which participants remained exposed. This tension (or exogenous incongruence) crystallised through employees’ tendency to enact traditional role patterns into the platform, which ran counter to the platform designers’ intention of enabling partic-

248  Research handbook on human resource management and disruptive technologies

ipants to ‘do different things’ (#1). For example, we observed that participants in supervisory positions primarily enacted a poster role and that lower-ranked participants mostly stuck to a worker role. Despite the intended affordance for participants to take on any role they wanted in the platform (that is, the platform enabled all participants to be posters and/or workers), participants were often reluctant to enact this affordance. For example, when reflecting on enacting the poster role, a participant told us: As I get more confidence, then I want to go outside of my comfort zone [by becoming a project poster]. But I’m like, ‘Whoa, I don’t want to go too far out because my boss is going to probably go: “No.”’ You just got to be a little bit careful. (#21)

Similarly, many posters showed reluctance to enact the worker role within the IOLP, citing managerial responsibilities and time constraints as reasons. For instance, a poster expressed: ‘a lot of times we don’t have the time [to be a worker in STRETCH]’ (#28). Along the same lines, a formerly active worker in STRETCH – after being promoted into a supervisory position – explained how ‘it is practically impossible to find any time in my schedule [to be a worker in STRETCH]’ (#14). These examples illustrate how employees enacted traditional role patterns in the IOLP despite the intended affordances for participants to move beyond these patterns, which only a minority of interviewees did. To break this pattern and help people move out of their comfort zone, STRETCH’s designers again sought to intervene in the platform’s material and social dimensions. However, the tension arising from employees’ hierarchical embeddedness proved hard to solve. Material changes were made, for instance, to make it easier to highlight their new experiences in their STRETCH profiles and to ‘aggregat[e] their experience along the way so that they don’t have to keep track of it in separate places and try to remember to combine and bring this all together’ (#20). While these changes helped employees to contemplate their gained experience and portray themselves differently within STRETCH, they did not directly target the forces that constrained employees from enacting new role patterns, such as the (perceived) resistance of their supervisors and formal and informal codes of conduct that worked against such behaviour. As a community manager acknowledged: ‘I don’t think [these material interventions] changed too much’ (#8). Besides these material changes, further interventions in the platform’s social dimension were made. These included efforts to help participants think differently about their activity in the platform, for instance, by organising departmental information sessions and presentations and providing direct support to employees. A community manager explained how these activities helped ‘break down assumptions about capabilities and the boxes that [participants] put [them]selves into’ (#27). She continued by emphasising the need to ‘get people to practice [new roles] in this safer way where they feel like it’s free labour and it’s not core to their work, [so] they can realise that they can [change]’ (#27). Other social interventions included the introduction of ‘champion participants’ – early adopters who worked with designers and community managers to ‘show others to live successful’ (#16) in the IOLP.

Platform work inside organisations  249

A designer explained how ‘it is essential to have those champions and then to even have a few within each government agency that can really go out and promote the platform’ (#4). While these social interventions typically eased employees’ concerns about their ability to try on roles that differed from their full-time roles, exogenous incongruences persisted. Lacking the reach to intervene in the platform’s hierarchical surroundings – of which employees remained an integral part – a designer acknowledged the limitations of their interventions: It’s just that mindset that’s really hard to shift … that traditional top-down thinking that we’ve always seen things run in the way things run. So, I think that’s going to take time [to] shift that culture and that idea. (#2)

The above examples show how participants’ enacted affordances can diverge from designers’ intended affordances, creating tension in the IOLP. However, unlike tension two, this tension originated from exogenous incongruences where organisational norms and routines outside the platform environment constrained participants in enacting designers’ intended affordances. While IOLP designers aimed to address this tension through interventions in the platform’s material and social dimensions, our data highlight the difficulty of addressing tensions arising from exogenous incongruences, which largely lie out of reach of such interventions.

CONCLUSION AND RECOMMENDATIONS FOR ACTION Our study examined how IOLPs create opportunities but also new organising tensions in the workplace. Labour platforms such as STRETCH effectively bring new labour concepts such as on-demand work, enhanced autonomy and platform mediation of work demand and supply (Meijerink and Keegan, 2019; Duggan et al., 2020; Vallas and Schor, 2020) inside organisations, even bureaucratic ones. Several studies have pointed to the emergent characteristics of IOLPs and their potential benefits, including IOLPs’ affordance of enhanced learning and self-development possibilities for workers (Rogiers et al., 2020, 2021) and their efficacy in mobilising the hidden potential of the workforce (Gantcheva et al., 2020; Schrage et al., 2020). We found, however, that IOLPs’ entrance into organisations also creates new tensions between the actors taking part in these novel labour platforms. Through an affordances and constraints perspective (Majchrzak and Markus, 2012; Volkoff and Strong, 2013; Leonardi and Vaast, 2017), we identified and documented the unfolding of three key tensions in our case context. Each exhibits a different type of tension in IOLPs: (a) arising from incongruently enacted affordances of platform participants (tension one), where the affordances enacted by one type of participant constrain others; (b) endogenous incongruences between designers’ intended affordances and participants’ enacted affordances (tension two), where participants experience designers’ intended affordances as constraining and as such, diverge from these intentions

250  Research handbook on human resource management and disruptive technologies

in their enacted affordances; and (c) exogenous incongruences between intended affordances and enacted affordances (tension three), where participants’ enactment of designers’ intentions is constrained by inertial forces originating outside the platform. We found that IOLPs can address these tensions by intervening in the platform’s material and/or social dimensions, and that tensions that lie out of reach of these interventions (i.e. because of exogenous incongruences) are likely to persist without a more coordinated organisational response. Perhaps one of the most significant insights from our study is that IOLPs do not so much compete with other platforms as with employees’ initial enthusiasm for the platform, their commitment toward their permanent jobs and business units and the ‘old’ norms and routines (Van Maanen and Schein, 1979; Jones, 1986; Miller and Jablin, 1991) to which they remain directly exposed. This competition dynamic creates unique tensions in IOLPs that, if unresolved, may set off participants, create labour retention problems and cause the IOLP to fall short of their promised potential. The identified organising tensions in our study show this risk. For instance, tension one illustrates the risk of incongruences between workers’ and posters’ enacted affordances (e.g. workers’ and posters’ enactment of their platform roles in ways that constrain each other). We showed how these incongruences cause tension among participants, resulting in disappointment and waning enthusiasm. Tension two further shows how endogenous incongruences (e.g. participants’ experience of designers’ market-logic as limiting) led participants to move away from the IOLP. This evolution posed a labour retention problem, which the platform designers tried to counter by building additional communal affordances into the IOLP that kept these communities tied to the platform. Finally, tension three, which remained unresolved because of exogenous incongruences between intended and enacted affordances (e.g. designers’ intention to afford new role patterns vs. participants’ dragging of traditional role patterns onto the IOLP), illustrates the risk for IOLPs to merely amplify extant role patterns – and fall short of their promise to break through the status quo. Rather than disrupting organisational routines and hierarchies, it turned out that such pre-existing routines and hierarchies equally disrupt the workings of IOLPs, if not well-managed. Besides its conceptual implications, our study also provides practical recommendations for business leaders looking to implement an IOLP in their organisation. First, while software vendors are increasingly offering new IOLP software platforms (Bersin, 2020), our findings show that implementing a successful IOLP goes beyond procuring and implementing a digital tool. For instance, we illustrated the critical role that platform designers play, not only in the initial design of IOLPs, but also in addressing emergent organising tensions. We showed how designers could do so in two distinct ways: through interventions in the platform’s material and social dimensions. Both interventions enable platform designers to alter the affordances and constraints dynamic behind emergent tensions in IOLPs, which is critical to sustaining the IOLPs’ balance in labour supply and demand. As a digital marketplace for short-term assignments, interventions in the social dimension are needed to com-

Platform work inside organisations  251

plement material interventions, tailored to the needs of platform participants in their different role(s) of project posters and project workers. Second, business leaders should realise that some IOLP tensions lie beyond the reach of platform designers’ interventions. In particular, tensions arising from exogenous incongruences between intended and enacted affordances are hard to manage by platform designers alone and require a more coordinated organisational response. In fact, IOLPs have to be embedded in a wider organisational model inscribed with material and social dimensions that reinforce the affordance of the IOLP by, for example, incentivising learning and development, proactivity, resource sharing and role flexibility. A critical role therein is reserved for business leaders and managers in key positions to craft proactive and coordinated interventions to IOLP tensions. For instance, leadership interventions could include setting up organisation-wide campaigns that signal top-down support of employees’ use of IOLPs and encourage employees to step outside of their comfort zone. Such campaigns could include organising awareness and training sessions that teach both employees and line managers about how to optimally use the IOLP. Finally, organisational leadership could set the example and take part in IOLPs themselves, for instance, by taking on the role of project worker – thereby showing that the IOLP embodies a new, encouraged way of working. Organisational leaders could also commit to intervene directly as a problem solver or use their authoritative capacity of leader in the ‘old’ organisation to facilitate and promote the adoption and use of IOLPs.

REFERENCES Altman, E. J., Kiron, D., Schwartz, J., and Jones, R. (2021). The future of work is through workforce ecosystems. MIT Sloan Management Review, 62(2), 1–4. Ashford, S. J., Caza, B. B., and Reid, E. M. (2018). From surviving to thriving in the gig economy: a research agenda for individuals in the new world of work. Research in Organizational Behavior, 38, 23–41. http://​dx​.doi​.org/​10​.1016/​j​.riob​.2018​.11​.001. Bersin, J. (2020). Talent Marketplace Platforms Explode Into View. The Josh Bersin Company, July 3. https://​joshbersin​.com/​2020/​07/​talent​-marketplace​-platforms​-explode​-into​-view. Boudreau, J. W. (2021). Engagement with perpetually deconstructed and reinvented work. In J. P. Meyer and B. Schneider (eds), A Research Agenda for Employee Engagement in a Changing World of Work (pp. 119–36). Edward Elgar Publishing. Boudreau, J. W., Jesuthasan, R., and Creelman, D. (2015). Lead the Work: Navigating a World beyond Employment. John Wiley & Sons. Campion, E. D., Caza, B. B., and Moss, S. E. (2020). Multiple jobholding: an integrative systematic review and future research agenda. Journal of Management, 46(1), 165–91. http://​ dx​.doi​.org/​10​.1177/​0149206319882756. Cecez-Kecmanovic, D., Galliers, R. D., Henfridsson, O., Newell, S., and Vidgen, R. (2014). The sociomateriality of information systems: current status, future directions. MIS Quarterly, 38(3), 809–30. Charmaz, K. (2006). Constructing Grounded Theory: A Practical Guide through Qualitative Analysis. Sage. Duggan, J., Sherman, U., Carbery, R., and McDonnell, A. (2020). Algorithmic management and app-work in the gig economy: a research agenda for employment relations and HRM.

252  Research handbook on human resource management and disruptive technologies

Human Resource Management Journal, 30(1), 114–32. https://​doi​.org/​10​.1111/​1748​-8583​ .12258. Faraj, S. and Azad, B. (2012). The materiality of technology: an affordance perspective. In P. M. Leonardi, B. A. Nardi, and J. Kallinikos (eds), Materiality and Organizing: Social Interaction in a Technological World (pp. 237–58). Oxford University Press. Gantcheva, I., Jones, R., Kearns-Manolatos, D., Schwartz, J., Lee, L., and Rawat, M. (2020). Activating the Internal Talent Marketplace: Accelerate Workforce Resilience, Agility and Capability, and Impact the Future of Work. Deloitte Workforce Transformation Practice. https://​www2​.deloitte​.com/​us/​en/​insights/​focus/​technology​-and​-the​-future​-of​-work/​ internal​-talent​-marketplace​.html. Gibson, J. J. (1979). The Ecological Approach to Visual Perception. Houghton Mifflin. Gioia, D. A., Corley, K. G., and Hamilton, A. L. (2013). Seeking qualitative rigor in inductive research: notes on the Gioia methodology. Organizational Research Methods, 16(1), 15–31. http://​dx​.doi​.org/​10​.1177/​1094428112452151. Glaser, B. G. and Strauss, A. L. (2017). Discovery of Grounded Theory: Strategies for Qualitative Research. Routledge. Grant, R. M. (1996). Toward a knowledge-based theory of the firm. Strategic Management Journal, 17(S2), 109–22. http://​dx​.doi​.org/​10​.1002/​smj​.4250171110. Hagiu, A. and Wright, J. (2015). Multi-sided platforms. International Journal of Industrial Organization, 43, 162–74. http://​dx​.doi​.org/​10​.1016/​j​.ijindorg​.2015​.03​.003. Jones, G. R. (1986). Socialization tactics, self-efficacy, and newcomers’ adjustments to organizations. Academy of Management Journal, 29(2), 262–79. https://​doi​.org/​10​.5465/​256188. Kiron, D., Schwartz, J., Jones, R., and Buckley, N. (2020). Create a crisis growth plan: start with opportunity marketplaces. MIT Sloan Management Review, 61(4), 1–5. Kuhn, K. M. and Maleki, A. (2017). Micro-entrepreneurs, dependent contractors, and instaserfs: understanding online labor platform workforces. Academy of Management Perspectives, 31(3), 183–200. https://​doi​.org/​10​.5465/​amp​.2015​.0111. Kuhn, K. M., Meijerink, J., and Keegan, A. (2021). Human resource management and the gig economy: Challenges and opportunities at the intersection between organizational HR decision-makers and digital labor platforms. Research in Personnel and Human Resources Management, 39, 1–46. Leonardi, P. M. (2013). Theoretical foundations for the study of sociomateriality. Information and Organization, 23(2), 59–76. https://​doi​.org/​10​.1016/​j​.infoandorg​.2013​.02​.002. Leonardi, P. M. and Vaast, E. (2017). Social media and their affordances for organizing: a review and agenda for research. Academy of Management Annals, 11(1), 150–88. https://​ doi​.org/​10​.5465/​annals​.2015​.0144. Maffie, M. D. (2020). The role of digital communities in organizing gig workers. Industrial Relations, 59(1), 123–49. https://​doi​.org/​10​.1111/​irel​.12251. Majchrzak, A. and Markus, M. L. (2012). Technology affordances and constraints in management information systems (MIS). In E. Kessler (ed.), Encyclopedia of Management Theory (pp. 832–5). Sage. Meijerink, J. and Keegan, A. (2019). Conceptualizing human resource management in the gig economy: toward a platform ecosystem perspective. Journal of Managerial Psychology, 34(4), 214–32. https://​doi​.org/​10​.1108/​JMP​-07​-2018​-0277. Meijerink, J., Keegan, A., and Bondarouk, T. (2021). Having their cake and eating it too? Online labor platforms and human resource management as a case of institutional complexity. International Journal of Human Resource Management, 32(19), 4016–52. https://​doi​ .org/​10​.1080/​09585192​.2020​.1867616. Miles, M. B. and Huberman, A. M. (1994). Qualitative Data Analysis: An Expanded Sourcebook (2nd ed.). Sage.

Platform work inside organisations  253

Miller, V. D. and Jablin, F. M. (1991). Information seeking during organizational entry: influences, tactics, and a model of the process. Academy of Management Review, 16(1), 92–120. https://​doi​.org/​10​.5465/​amr​.1991​.4278997. Nahapiet, J. and Ghoshal, S. (1998). Social capital, intellectual capital, and the organizational advantage. Academy of Management Review, 23(2), 242–66. https://​doi​.org/​10​.5465/​amr​ .1998​.533225. Nelson, R. R. and Winter, S. G. (1982). An Evolutionary Theory of Economic Change. Harvard University Press. Orlikowski, W. J. (2010). The sociomateriality of organisational life: considering technology in management research. Cambridge Journal of Economics, 34(1), 125–41. https://​doi​.org/​ 10​.1093/​cje/​bep058. Rogiers, P., Viaene, S., and Leysen, J. (2020). The digital future of internal staffing: a vision for transformational electronic human resource management. Intelligent Systems in Accounting, Finance and Management, 27(4), 182–96. https://​doi​.org/​10​.1002/​isaf​.1481. Schrage, M., Schwartz, J., Kiron, D., Jones, R., and Buckley, N. (2020). Opportunity marketplaces: aligning workforce investment and value creation in the digital enterprise. MIT Sloan Management Review. https://​sloanreview​.mit​.edu/​projects/​opportunity​-marketplaces. Sessions, H., Nahrgang, J. D., Vaulont, M., Williams, R., and Bartels, A. L. (2021). Do the hustle! Empowerment from side-hustles and its effects on full-time work performance. Academy of Management Journal, 64(1), 235–64. https://​doi​.org/​10​.5465/​amj​.2018​.0164. Smith, W. K. and Lewis, M. W. (2011). Toward a theory of paradox: a dynamic equilibrium model of organizing. Academy of Management Review, 36(2), 381–403. https://​doi​.org/​10​ .5465/​amr​.2009​.0223. Spreitzer, G. M., Cameron, L., and Garrett, L. (2017). Alternative work arrangements: two images of the new world of work. Annual Review of Organizational Psychology and Organizational Behavior, 4, 473–99. https://​doi​.org/​10​.1146/​annurev​-orgpsych​-032516​ -113332. Stake, R. E. (2000). Case studies. In N. K. Denzin and Y. S. Lincoln (eds), Handbook of Qualitative Research (2nd ed., pp. 435–54). Sage. Stanford, J. (2017). The resurgence of gig work: historical and theoretical perspectives. The Economic and Labour Relations Review, 28(3), 382–401. https://​doi​.org/​10​.1177/​ 1035304617724303. Strauss, A. and Corbin, J. M. (1997). Grounded Theory in Practice. Sage. Vallas, S. and Schor, J. B. (2020). What do platforms do? Understanding the gig economy. Annual Review of Sociology, 46, 273–94. https://​doi​.org/​10​.1146/​annurev​-soc​-121919​ -054857. Van Maanen, J. E. and Schein, E. H. (1979). Toward a theory of organizational socialization. Research in Organizational Behavior, 1, 209–64. Volkoff, O. and Strong, D. M. (2013). Critical realism and affordances: theorizing IT-associated organizational change processes. MIS Quarterly, 37(3), 819–34. Weller, I., Hymer, C. B., Nyberg, A. J., and Ebert, J. (2019). How matching creates value: cogs and wheels for human capital resources research. Academy of Management Annals, 13(1), 188–214. https://​doi​.org/​10​.5465/​annals​.2016​.0117.

18. Empowering or taking over? A job design perspective on the effects of cobots’ introduction in the manufacturing industry Emanuela Shaba, Alessandra Lazazzara, Luca Solari and Antonella Delle Fave

INTRODUCTION This chapter is focused on the disruptive effects that the introduction of collaborative robots (aka cobots; El Zaatari et al., 2019) in the manufacturing context has on job design. Compared to traditional industrial robots, designed to operate in isolation and with minimal physical interaction with humans (Arviv et al., 2016), this new generation of collaborative robots is designed to share the workspace with human operators (Kadir et al., 2018; El Zaatari et al., 2019). In aiming to distinguish cobots from traditional robots, Cohen et al. (2022) focused on key structural and operational features of cobots, such as mobility, i.e. the ability to easily move around in the production plant; intelligence, i.e. the capability to interact with operators by enabling gesture recognition, speech recognition and anticipating operator moves; connectivity, i.e. human-cobot and cobot-cobot communication; actuation, i.e. the ability to develop safe and dynamic trajectories; and human-centricity, the support provided to the human operator from the physical, mental and psychosocial point of view. Through these features, cobots not only bring productivity and efficiency gains to companies, but also support organizations to operate under dynamic and unique production conditions based on principles of flexibility, customer demand and product variation and quality (Vasic and Billard, 2013). Moreover, as a cobot safely shares the workspace with human operators, it can be quickly reprogrammed and easily applied in several industrial scenarios (Sherwani et al., 2020), whereby, they are especially advantageous and most commonly used in assembly tasks, where the high payload and repeatability characterizing traditional robotic systems need to be combined with the dexterity and flexibility of a human operator (Matheson et al., 2019). Given such benefits, cobots are expected to undergo the fastest growth across several sectors such as electronics, automotive, logistics, packaging and assembling and machine tooling applications. In fact, the global market for cobots is predicted to grow by 42.80% by 2028, experiencing particular growth in small and medium enterprises (SMEs).1 254

Empowering or taking over?  255

The interaction between humans and robots (HRI) in the industrial domain relies on decades of research (Villani et al., 2018). Existing studies in HRI research have mostly focused on increasing functional and safety requirements, thanks to the design of increasingly refined technologies and interaction patterns (Kim et al., 2019; Gasparre and Tirabeni, 2020). Thereby, many HRI studies have grappled with aspects such as aesthetics (Gibson et al., 2021), affective interaction, embodied action, mobility (El Zaatari et al., 2019), or even tried to deepen the psychological and emotional effects of that interaction within experimental settings (Jerčić et al., 2019). With the new generation of collaborative robots, as humans and machines are no longer working separated by cages, aspects such as proximity, coordination and harmony in the operating conditions have represented new challenges for HRI research (Welfare et al., 2019). Research shows that such aspects are highly correlated with usability judgments and have the potential not only to increase cobots’ acceptance in the workplace, but also an operator’s motivation to interact (Pollak et al., 2020). However, given that cobots work jointly with humans and they are sharing routine and nonroutine manual and cognitive tasks, human-cobot collaboration (HCC) research, albeit in its infancy, has endeavored to explore the implications that cobot introduction has on shared tasks at the level of intersection (Kadir et al., 2018; Smids et al., 2019; Pollak et al., 2020; Berkers et al., 2022). According to Parker and Grote (2020), understanding the implications that new technology (such as cobot introduction) has on shared tasks, and hence on aspects of individual work, can drastically affect employee responses such as their wellbeing and performance. Based on such considerations, the objective of this chapter is to theoretically explore the implications that human-cobot collaboration has on job characteristics, as categorized from the ‘job demands’ (Bakker and Demerouti, 2007) and ‘job resources’ work characteristics model. Job demands and resources capture key aspects of work design from a range of theories, including also the dominant job characteristics model, which focuses on aspects (e.g. skills’ use, variety in one’s work, etc.) that have subsequent effects on employee motivation (Parker and Grote, 2020), and hence are the basis of employee outcomes (Morgeson and Humphrey, 2006; Parker et al., 2017b; etc.). Despite the scarce existing evidence, results of this study show that cobot introduction has both positive and negative effects on multiple aspects of work, bringing attention to the paradoxes that accompany job design aspects. Based on the results of the study, we argue that exploring the effects that cobot introduction has on the motivating aspects of the job may help research explain the effects that human-robot collaboration has on employee response, an important focus of HRI research. Thereby, we conclude that proactive efforts to shape work design, alongside human-centered design of cobot technology, are likely to generate not only performance benefits, but also better meet human competencies, needs and values.

256  Research handbook on human resource management and disruptive technologies

The Emerging of Collaborative Robots: New Directions in Human-Robot Interaction Research The goal of the introduction of such disruptive technology in the manufacturing industry is the creation of teams of humans and cobots effectively interacting and taking advantage of each team member’s skills. Such interaction has raised researchers’ interest on a number of issues that have become increasingly discussed in human-robot interaction (HRI) research. For years, HRI research has adopted an anthropomorphic or cognitive approach towards interaction design, in order to grapple with aspects such as aesthetics, affective interaction, embodied action, mobility and situated aspects of human-robot shared activity (El Zaatari et al., 2019). The introduction of cobot technology has presented new challenges for HRI research. Even if the main focus of cobot design has mostly been on building safe and functional aspects for human-robot collaboration (El Zaatari et al., 2019; Kim et al., 2019), HRI literature has also dealt with other aspects that challenge such collaboration. The study conducted by Vanzo et al. (2019) showed, for example, how a human’s collaboration attitude is in fact closely related to proxemics, highlighting how the relative position of the interactive partners affects such collaboration. Hoffman (2019) provides tools for evaluating ‘collaborative fluency’ in order to build a well synchronized, coordinated and joined-up sequence of activities between human-cobot team members. Thought has also been given to the appearance of cobots, as a cobot supposed to collaborate with humans should not look threatening but friendly (Gibson et al., 2021). Other studies have been mostly focused on the psychological impact of the human-cobot collaboration, as it can produce aversive and negative attitudes, which can damage the motivation of the human to interact with collaborative robots (Ahmad et al., 2017; Jerčić et al., 2019; Belhassein et al., 2022). However, as collaborative robots work side by side with humans, and as their adoption has become ever more pervasive, many authors (Carissoli et al., 2016; Smids et al., 2019; Welfare et al., 2019; Pollak et al., 2020) have recently focused on joint action at the level of task intersection, arguing that the collaboration has important implications for aspects of human work, which in turn is bound to have implications for an individuals’ cognitive, social and identity development in human-robot interaction. New Collaborative Technologies: A Threat or an Opportunity for Workers? In the Industry 4.0 scenario, as automation and digitalized machines/cobots that work alongside humans and depend on each other will intensify, it is necessary to understand the implications that technology has for work. There is a sound body of knowledge on the impact of Industry 4.0 technology on work (e.g. Acemoglu and Restrepo, 2018; Autor and Salomons, 2018), mostly focused on the implications that such technology has on the future of jobs. For example, many authors have tackled the social impacts of 4.0 technology and robotics, tending to glorify their

Empowering or taking over?  257

effects on work; these authors argue that, far from reducing work, such technology is likely to improve it (Huws, 2014; Spencer, 2018) by extending human freedom (Srnicek and Williams, 2015) and yielding more liberating options for the meaning of work (Neufeind et al., 2018). By contrast, other perspectives focus on job loss (van Wynsberghe and Comes, 2020), with studies that predict in-work poverty, arguing that new jobs are likely to become more insecure and less rewarding and careers more fragmented. Schumpeter (2015), and Moore and Robinson (2016) have argued that jobs are characterized by a low number of simple activities, with little or no room for maneuver, in a way that can be addressed to as ‘Neo Taylorism’. Fleming (2019), while holding a ‘bounded automation’ perspective, has argued that technological innovations do not simply unfurl according to their own endogenous potential, elucidating that a robot probably will not steal jobs from humans, but that’s no cause for celebration because the jobs that will proliferate in the ‘second machine age’ are considerably poorer in terms of skill, responsibility and pay. Fueled by the acceleration of digitization during the COVID-19 crisis, such dystopian sentiments on the future of work have found more ground (Schlogl et al., 2021). Rather than solely speculating about which jobs will vanish, Parker and Grote (2020) argue that research should address the urgent and prevalent matter of how tasks might best be shared between humans and machines, and what might be the consequences of different choices in this respect. Understanding which tasks people covet or dislike the most while working side by side with collaborative robots becomes thus necessary to disentangle the paradoxes that cobot introduction brings to aspects of work. Implications of New Technologies for Job Design Job design, with a long and rich tradition in industrial and organizational psychology, refers to the content and organization of work tasks, activities, relationships and responsibilities (Parker, 2014). As early writings designed work to maximize efficiency, jobs tended to result in decreased employee satisfaction, increased turnover and absenteeism and difficulties in managing employees in simplified jobs (Morgeson and Humphrey, 2006). Reacting to this perspective, researchers developed theories focusing on the motivating features of work. In particular, the motivation-hygiene theory (Hackman and Oldham, 1976, 1980) became a very influential approach over the years. According to the job characteristic model (Hackman and Oldham, 1976), for each job it is possible to analyze work characteristics, including but not limited to autonomy, variety, identity, significance and feedback, as well as knowledge characteristics such as complexity, information processing, problem-solving, skill variety and specialization (Parker et al., 2001), and physical characteristics, for instance ergonomics, physical demands, work conditions and equipment use. In order to be able to address other aspects of work, other authors have proposed different job design models, such as job crafting (Bakker et al., 2012), a process through which workers reinvent their job. However, job crafting strictly depends on the contextual conditions (e.g. organizational climate, technology, managerial style), which are

258  Research handbook on human resource management and disruptive technologies

central in linking individual reactions to different job design components (Lazazzara et al., 2020). While different models encompass different job and work characteristics, the consensus is that work design leads to critical psychological states, which in turn are the basis for individual and organizational consequences (Morgeson and Humphrey, 2006). In fact, as technological advances bring challenges and opportunities for organizations to the quantity and nature of work (e.g. Morgeson and Humphrey, 2008; Parker et al., 2017a), there is a sound body of knowledge on the impact of technology on work design (e.g. Morgeson and Humphrey, 2006; Parker et al., 2017a). Parker and Grote (2020) summarize examples of how AI and digital technology can affect work design, both positively and negatively, according to the following broad categories of work characteristics: (I) job autonomy and control, including decision-making over work processes and methods; (II) skill variety and use, as new technologies can provide greater opportunity for individuals to engage in skilled and meaningful tasks; (III) job feedback and related work characteristics; and (IV) job demands. For example, physical demands can change with technology as heavy manual work is replaced by automation, but on the other side technology can require higher cognitive skills. Hence, as new technology does not affect just one single work characteristic, but multiple aspects of work simultaneously, the introduction of cobots in the workplace calls into question the many implications it has for job design.

IMPACTED JOB CHARACTERISTICS IN HUMAN-ROBOT COLLABORATION: A PARADOXICAL VIEW With the introduction of cobots, the academic discussion has been centered mostly on the functional aspects of such collaboration and also on its implications on the future of jobs. In addition, literature that deals with the changing nature of work at the level of task intersection is sparse and it presents many contradictions. In the following pages we endeavor to provide an overview of such contradictions, pertaining to the present and existing risks and opportunities related to different aspects of job characteristics. Holding a job design perspective, in Table 18.1 we summarize different facets (related to routine and nonroutine manual and cognitive tasks) of work characteristics, such as autonomy, task variety, physical and cognitive demands, social relationships and skill development, that get simultaneously imbued with various positive and negative effects. Lack of Autonomy or Opportunity for Proactive Behavior? According to Smids et al. (2019), the way humans collaborate with robots has given rise to more standardization and monitoring of work, which appears to be a threat to the worker’s autonomy (Cascio and Montealegre, 2016; Lanzing, 2016). In fact, some robotic applications lead to an increased level of standardization in the workplace;

Empowering or taking over?  259

Table 18.1

Summary of the effect of cobot introduction on jobs according to the literature

Job Characteristics

Opportunities for Workers

Threats for Workers

Nature of Work

Cobots are taking over some boring or

Cobots are taking over processes such as:

Cognitive Demands

tedious tasks (and not the most difficult constant searching, thinking and checking ones), hence human workers focus

of orders, thus humans are experiencing

on more meaningful-seeming tasks

a reduction in the number of nonroutine

(Welfare et al., 2019)

cognitive tasks demands (Wang et al., 2020; Berkers et al., 2022)

Cognitive tasks such as short-term

With machine learning cobots, some of the

memory, concentration improvement

most challenging tasks will be taken over

and error reduction might benefit from

from cobots, as they have higher accuracy

cobot support (Senders et al., 2019;

rates (Esteva et al., 2017; Senders et al.,

Smids et al., 2019)

2019; Smids et al., 2019)

Cobots lead to intensification of

Pace of work is imposed, thereby it

cognitive tasks that make work more

increases mental workload due to temporal

challenging and meaningful (Berkers

demands (Pham et al., 2018; Charles

et al., 2022)

and Nixon, 2019; Argyle et al., 2021); With cobot-assessment of MW (mental workload) there is increase in the degree of

Physical Demands

work disruption (Argyle et al., 2021) Cobots performing routine heavy

Cobots have allowed also the automation of

burden tasks (Autor, 2015; Bragança

some nonroutine manual tasks, especially

et al., 2019; El Zaatari et al., 2019;

those requiring situational adaptability,

Mühlemeyer, 2020; Pollak et al.,

visual and language recognition (Arntzt et

2020), offloading the heavy burden of

al., 2021)

repetitive, straining tasks

Cobot technological malfunctions generate pressure on lead times in logistics; In turn employees have to work even harder to make up for lost time later (Jaehrling et al., 2018; Berkers et al., 2022)

Task Variety

As the cobot is performing repetitive

With the cobot introduction, tasks such as

and physically demanding tasks (El

planning a route, walking and looking up

Zaatari et al., 2019), the worker is

the product in the warehouse dropped from

taking over the more flexible and high

the job description; Reduced number of

variation parts of the tasks (Kadir et

tasks (Berkers et al., 2022)

al., 2018)

260  Research handbook on human resource management and disruptive technologies Job Characteristics

Opportunities for Workers

Threats for Workers

Autonomy

Employees use their capacities

More standardization of work processes

for understanding, judgement and

(Sanders et al., 2016; Smids et al., 2019;

decision-making more for overseeing

Welfare et al., 2019), greater threat to

the cobot (Wingfield, 2017; Berkers et

decision-making (Cascio and Montealegre,

al., 2022)

2016; Lanzing, 2016); Fewer opportunities for job crafting (Smids et al., 2019)

Employees may exert autonomy by

Strong emphasis on the cobot’s

proactively changing their work after

technological functionality has impacted,

cobots are implemented (Berkers et

i.e., work methods and scheduling

al., 2022)

autonomy (Berkers et al., 2022) The built-in AI, hence, reliance on algorithm-based for decision-making poses a threat to humans’ decision-making (Burrell, 2016)

Skill Development

As there is a shift toward complex

Risk of losing manual skills or other skills

scenarios that require a more

(Riek, 2017; Welfare et al., 2019); Dangers

knowledge-based approach to the work, of deskilling due to Cobot machine learning cobot introduction requires workers to

techniques as the need to extensively train

acquire new and additionally complex

humans might disappear (Smids et al.,

skills (Bragança et al., 2019; El Zaatari 2019). et al., 2019) Working with cobots opens up the

Reduced opportunities for trainees to

opportunity for interesting tasks and

engage in challenging tasks, in knowledge

to learn new skills in ways similar to

and the skill to perform such tasks;

advanced manufacturing (Smids et al.,

Diminished need to exercise skills generates

2019; Berkers et al., 2022)

danger of deskilling (Beane, 2018; Smids et al., 2019)

this requires working according to a very strict protocol, which leaves little room for human creativity, judgment and decision-making. For the same reasons, workers’ opportunities to engage in job crafting may be severely restricted. Their tasks and work environment may be so tightly structured by the cobots that there is little room for restructuring, therefore undermining worker’s autonomy. Berkers et al. (2022) have highlighted that the strong emphasis on the cobots’ technological functionality has resulted in design decisions that have reduced employees’ autonomy in working with collaborative robots. For example, collaborative robots negatively affected the autonomy of order pickers and packers, as employees became dependent on how fast and when robots worked, a process that impacted work methods and scheduling autonomy. This in turn resulted in frustration, as it decreased the operator’s work pace. It is in fact also evident from other studies that the length of waiting time (as humans have to wait for the cobot and thus stop work) relates negatively to the operator’s scheduling autonomy (Riek, 2017; Welfare et al., 2019). Another threat to human autonomy has to do with the worker’s understanding of the job. Collaborative robots incorporate artificial intelligence, which often involves machine learning and artificial neural networks. For most people, these AI techniques are hard to under-

Empowering or taking over?  261

stand beyond the surface level and therefore difficult to control and explain to others when needed. This phenomenon is commonly referred to as the opacity of artificially intelligent systems (Burrell, 2016), which may lead to feelings of alienation and diminished human autonomy. On the other side, some studies have brought examples of workplaces designed in ways that allow human-cobot collaboration, so that both parties team up while leaving room for autonomous human action. For example, Wingfield (2017) highlights how an Amazon warehouse employee certainly uses the capacities for understanding, judgment and decision-making more when overseeing robots than when stacking plastic bins. Therefore, human responsibility and autonomy have increased upon cobotization of the workplace. Berkers et al. (2022) show how employees exerted their autonomy by proactively changing their work after the implementation of cobots. For example, as the cobot assisted the picker with locating the right products, or during packing products and putting them on the conveyor belt, order packers proactively developed unique strategies of feeding the ‘order packing robot’ packages in a particular order to avoid malfunctions. The author argues that in one of the observed warehouses work was intentionally (re)designed to include such autonomy in the collaboration layout. Task Variety or ‘Left-Over’ Tasks? The way collaborative robots are implemented can negatively affect the amount of and variety in tasks. In the study of Berkers et al. (2022), the amount that order packers and order pickers handled daily got reduced. As a result, tasks such as planning a route, walking and looking up the product in the warehouse, were dropped from the order picker’s job description. Berkers et al. (2022) show that the reduction in tasks was rarely compensated with new tasks. Most complex tasks associated with robotization, such as handling errors or overseeing a cobot, were rarely assigned to order pickers or packers. This study has thus reported that work has become simpler and more monotonous, as cobots have taken over tasks (cobot automatically brings the products to the operator). The study from Berkers et al. (2022) indicated that such an aspect was negatively appraised by both employees and managers, as ‘the more robots or conveyors do, the less people do, and the simpler work becomes, making it hard to keep people satisfied because work becomes monotonous’. The predominantly negative effect of cobotization on task variety may also be explained by the work design approach of all warehouses focused on technological functionality. On the other side, there are studies showing that, where the introduction of cobots has improved the workflow and continuous production, the cobot is performing the repetitive and physically demanding tasks, while the worker is taking care of the more flexible and high variation parts of the tasks; this solution leaves workers with more time to spend on other and new potentially value-creating activities (Kadir et al., 2018). As reported in the theoretical background section of this chapter, the decision makers as well as workers consider this as one of the greatest benefits that derive from the implementation of the cobots. Kadir et al. (2018) in their empirical

262  Research handbook on human resource management and disruptive technologies

work show how workers were able to occupy new functions and take on new roles and responsibilities, such as: preparing products for storage, cleaning and servicing machines, programming cobots and other machines, designing new collaborative work with cobots, management of small internal projects, quality control, production planning, making technical drawings and performing other manual tasks in the production. However, the authors note that all these tasks were not officially part of the workers’ job descriptions. Cognitive Demands: Pursuing a Purpose or Losing It, while Teaming Up with Cobots? The outcomes related to the implications that cobot introduction have on cognitive demands are mixed. Significant reductions in the routine manual and cognitive tasks have been possible, as such tasks are performed under well-understood procedures that can be easily programmed and hence performed by cobots at an economically viable cost (Pollak et al., 2020). Research also shows that cobot introduction has generated a reduction of some nonroutine cognitive demands, such as constant searching, thinking and checking of orders. The study from Berkers et al. (2022) provides exemplary cases of cobots taking over cognitive tasks: light to pick or voice to pick robots assist with locating the product; the checking cobot weighs boxes and compares actual and expected weight to detect errors. In addition, Smids et al. (2019) argue that through machine learning cobots will take over some more challenging tasks, and hence workers might feel that they serve less of a purpose, as robotic systems that employ machine learning techniques have accuracy rates comparable to, and often higher than, humans (Esteva et al., 2017; Senders et al., 2019; Wang et al., 2020). However, other studies show that if cobots help or assist with tasks, rather than fully taking them over, or if they take over some boring or tedious tasks leaving to workers the most difficult and challenging ones, humans can still justifiably feel that they have a clear purpose, including being able to focus on more meaningful-seeming tasks. Smids et al. (2019) and Senders et al. (2019) argue to that end that if human workers perceive themselves as teaming up with robots, they may focus on achieving better outcomes together with the robots. These studies suggest how cognitive tasks such as short-term memory, concentration improvement and error reduction, might benefit from cobot support. Moreover, Welfare et al. (2019) highlight that the reduction of waiting time as a technological attribute has mostly positive effects, as people would have more time to work or to complete more interesting tasks. Through interviews, Welfare et al. propose that a cobot could be used for ‘grabbing the [materials] when they run out because then that’s one less thing for workers to have to go out and search for and come back’. This would reduce time spent searching for materials and allow a worker to stay focused on the required task (Welfare et al., 2019). Also, Berkers et al. (2022) argue that order pickers and order packers experience an intensification of cognitive tasks as they spend more time focusing on not making mistakes, which has made their work more challenging and meaningful.

Empowering or taking over?  263

Offloading Physical Requirements Related to Physical Strength, Endurance, Exercise and Work Activity, and Onloading Mental Burden? Cobots, in fact, are especially advantageous and most commonly used in assembly tasks, where the high payload and repeatability characterizing traditional robotic systems need to be combined with the dexterity and flexibility of a human operator (Matheson et al., 2019; Berkers et al., 2022). In fact, Pollak et al. (2020) argue that as many tasks were straining due to their repetitive nature, ‘the cobot has partly taken over the repetitive work of humans, which works very well for operators’. However, cobots also allow for the automation of some nonroutine manual tasks, especially those requiring situational adaptability and visual and language recognition (Arntzt et al., 2021). These aspects of work design are essential for robot operators, and they are related to nonroutine cognitive analytical and interpersonal tasks. Their performance relies on creativity, originality as well as social perceptiveness and empathic response to a human counterpart, which remains a highly demanding area to automate (Arntz et al., 2021). On the other side, workers who retain their jobs alongside cobots might not always see their conditions improve, as some studies report an increased level of work intensity (Pham et al., 2018). For instance, in Amazon warehouses, robots have been introduced on a massive scale over the past few years (Selby, 2017; Pham et al., 2018); because the cobots are so fast and so consistent, their pace can be set arbitrarily and it is, in fact, imposed on the workers (Argyle et al., 2021), leading to an increase of mental workload during human-robot collaboration. Mental workload is a property emerging from the relationships between physical and cognitive task demands, additionally affected by factors including temporal demand, individual background experience and environmental factors (Charles and Nixon, 2019). Berkers et al. (2022) argue that when cobots stopped working due to technological malfunctions, employees were forced to stop and wait. Due to the increased pressure to reduce lead times (e.g. same-day delivery (Jaehrling et al., 2018), employees had to work even harder to make up for lost time later or switch back to working manually. Upskilling or Downskilling: Is Self-Development under Threat? The introduction of cobots in the workplace impacts on human skill development in multiple ways; this impact needs to be assessed on a case-by-case basis. Like any machines introduced into the production process, cobots have contrasting effects on skills. On one side, if cobots take over more complex tasks from human workers, several human skills may become obsolete. There is the risk of losing manual or other skills with increased monitoring and standardization of tasks (Sanders et al., 2016; Riek, 2017; Welfare et al., 2019). To that end, Beane (2018) in his work showed how robotic technologies – designed for efficiency – created a finer-grained division of work, thereby restricting roles to more routine tasks, and radically reducing operator’s time for learning. While the study from Beane (2018) is an example of impaired feedback, learning and skill use, she showed that robotic technology can reduce the

264  Research handbook on human resource management and disruptive technologies

opportunities for trainees to engage in challenging tasks ‘at the edge of their skills and competence’. The lack of development and exercise of skills in this case will no longer be a source of meaningfulness for human workers, and their job will be less conducive to self-realization. In addition, the dangers of deskilling due to reliance on cobot technology are real if machine learning techniques become systematically better than humans. To that end, the need to extensively train humans might disappear (Smids et al., 2019). As a consequence, together with a diminished need to exercise one’s skills, one’s work-related growth and self-development will probably suffer (Smids et al., 2019). However, on the other side, considering the pervasive shift toward complex work scenarios that require a more knowledge-based approach to work, cobots might equally well have the opposite impact on skills. Thereby, they can enhance the need for workers to both maintain their skills and acquire new and additional complex ones, in particular, when workers have to face complex productive processes where new sets of skills should be developed (Bragança et al., 2019). In this case, humans have to acquire further complex skills to operate the cobot technology (e.g. the development of intuitive cobot programming allows nonexpert operators to create and alter robot programs) (El Zaatari et al., 2019). Smids et al. (2019) further show that, in order to work together with the warehouse robots, the individual has to maintain all the traditional skills and he/she has to acquire new complex skills needed for overseeing and handling cobot technology. According to Berkers et al. (2022), working with cobots gave some order pickers and packers new, interesting tasks and the opportunity to learn new skills in ways similar to advanced manufacturing. However, these opportunities were limited, indicating that the potential upskilling (Parker and Grote, 2020), at least for these employees, rarely took place.

CONCLUSION AND RECOMMENDATIONS FOR ACTION In taking a closer look at the results, two reflections arise. First, the opposing logics approach shows that it is important to understand the reason why cobot adoption has been accompanied by highly contradictory results. How can the cobot technology be disempowering and simultaneously empowering at the level of specific tasks? In their milestone paper, Edwards and Ramirez (2016) suggest that focusing on different dimensions of 4.0 technologies, in terms of their intended and unintended effects; direct and indirect effects, etc., could help future research explain how homogeneous outcomes (such as the standardization of work processes) become common across organizations, such that they shape the wider contours of work, while job characteristics variables present more heterogeneous behavior. Hence, in understanding the above paradoxes and contradictions, and as cobot presence becomes more pervasive, there is a need to explore different dimensions to understand how their adoption, implementation and use shape work and organizations. Based on the results described in the previous pages, the contributions of this study are twofold. First, it contributes to enriching work design theory, by presenting all

Empowering or taking over?  265

the contradictions that characterize job characteristics in relation to human-robot collaboration and also by providing specific examples of the work processes in which such paradoxical results are rooted. Secondly, the study shows that the integration of both social and technical systems should be considered by work designers right at the start of the technological design process, to avoid a techno-centric approach to work. Berkers et al. (2022) in their study argue that, as cobots were designed to take over certain tasks, their negative effects were not unexpected. In fact, as the authors argue, in aiming for work simplification, the observed organization created low-quality jobs around the so called ‘left-over tasks’ that cobots could not perform. Their study is a perfect example of how the social system was intentionally subordinate to the technical system while designing human-cobot collaboration, and that the effects of technology on work design can vary depending on various factors, such as how the technology is implemented and used. Hence, it is important to design workplaces where cobot flexibility could promote a functional match between job task challenges and workers’ skills, in order to support human motivation and engagement. In fact, this review is part of a Horizon 2020 research project, namely the Mindbot2 project, that is aimed at designing a new type of cobot (called Mindbot) able to interact with the human operator in ways that not only promote a functional match between job task demands and workers’ skills, but also make the task execution challenging, motivating and not frustrating, while respecting workers’ autonomy. Such an approach to cobot introduction is a demonstration that organizations can choose to ‘disobey the technological imperative’ and that technologies can be designed in ways that tap and enhance performance and develop human potentialities and, consequently, their wellbeing. In addition, as introduction of cobots brings disruptive change for work and related performance, this review presents original and interesting implications for human resource practitioners. In cobot-based manufacturing scenarios, workers are at risk for being assigned monotonous and alienating tasks where just low skills are required, or too challenging tasks where a high production ratio is expected. As work can be a source of self-esteem, social recognition and recognition of one’s skills and accomplishments (Fullagar and Delle Fave, 2017; Delle Fave and Massimini, 2005), the ongoing task for HRM is to pay attention to the challenging aspects of work design, to fit with present and future technological, economic objectives and social needs (human development).

ACKNOWLEDGMENTS This project has received funding from the European Union’s Horizon 2020 research and innovation program, under grant agreement No 847926.

266  Research handbook on human resource management and disruptive technologies

NOTES 1. Collaborative Robots Market Global Size, Share, Forecast 2022–2028, Vantage Market Research. 2. Further information on the Mindbot Project: https://​www​.mindbot​.eu/​.

REFERENCES Acemoglu, D. and Restrepo, P. (2018). Robots and jobs: evidence from US labor markets. NBER Working Paper. Ahmad, M., Mubin, O., and Orlando, J. (2017). A systematic review of adaptivity in human-robot interaction. Multimodal Technologies and Interaction, 1(3), 14. Argyle, E. M., Marinescua, A., Wilson, M. L., Lawsona, G., and Sharples, S. (2021). Physiological indicators of task demand, fatigue, and cognition in future digital manufacturing environments. International Journal of Human-Computer Studies, 145, 102522. Arntz, A., Rijkeboer, M., Chan, E., Fassbinder, E., Karaosmanoglu, A., Lee, C. W., and Panzeri, M. (2021). Towards a reformulated theory underlying schema therapy: position paper of an international workgroup. Cognitive Therapy and Research, 45, 1007–20. https://​ doi​.org/​10​.1007/​s10608​-021​-10209​-5. Arviv, K., Stern, H., and Edan, Y. (2016). Collaborative reinforcement learning for a two-robot job transfer flow-shop scheduling problem. International Journal of Production Research, 54(4), 1196–209. https://​doi​.org/​10​.1080/​00207543​.2015​.1057297. Autor, D. H. (2015). Why are there still so many jobs? The history and future of workplace automation. Journal of Economic Perspectives, 29(3), 3-30. Autor, D. H., and Salomons, A. (2018). Is Automation Labor-Displacing? Productivity Growth, Employment, and the Labor Share. National Bureau of Economic Research, Cambridge. Bakker, A. B. and Demerouti, E. (2007). The job demands-resources model: state of the art. Journal of Managerial Psychology, 22(3), 309–28. Bakker, A. B., Tims, M., and Derks, D. (2012). Proactive personality and job performance: the role of job crafting and work engagement. Human Relations, 65(10), 1359–78. https://​doi​ .org/​10​.1177/​0018726712453471. Beane, M. (2018). Shadow learning: building robotic surgical skill when approved means fail. Administrative Science Quarterly, 64(1), 87–123. https://​doi​.org/​10​.1177/​ 0001839217751692. Belhassein, K., Fern, V., Castro, F., Mayima, A., Clodic, A., Guidetti, M., Pacherie, E., Alami, R., and Cochet, H. (2022). Addressing joint action challenges in HRI: insights from psychology and philosophy. Acta Psychologica, 222. Berkers, H. A., Rispens, S., and Le Blanc, L. M. (2022). The role of robotization in work design: a comparative case study among logistic warehouses. The International Journal of Human Resource Management. https://​doi​.org/​10​.1080/​09585192​.2022​.2043925. Bragança, S., Costa, E., Castellucci, I., and Arezes, P. M. (2019). A brief overview of the use of collaborative robots in Industry 4.0: Human role and safety. Studies in Systems, Decision and Control. Burrell, J. (2016). How the machine ‘thinks’: understanding opacity in machine learning algorithms. Big Data & Society, 3(1). https://​doi​.org/​10​.1177/​2053951715622512. Carissoli, C., Villani, D., and Riva, G. (2016). Does a meditation protocol supported by a mobile application help people reduce stress? Suggestions from a controlled pragmatic trial. Cyberpsychology, Behavior, and Social Networking, 18(1), 46–53. 10.1089/ cyber.2014.0062. PMID: 25584730.

Empowering or taking over?  267

Cascio, W. F. and Montealegre, R. (2016). How technology is changing work and organizations. Annual Review of Organizational Psychology and Organizational Behavior, 3(1), 349–75. https://​doi​.org/​10​.1146/​annurev​-orgpsych​-041015​-062352. Charles, R. L. and Nixon, J. (2019). Measuring mental workload using physiological measures: a systematic review. Applied Ergonomics, 74, 221–32. Cohen, Y., Shoval, S., Faccio, M., and Minto, R. (2022). Deploying cobots in collaborative systems: major considerations and productivity analysis. International Journal of Production Research, 60(6), 1815–31. https://​doi​.org/​10​.1080/​00207543​.2020​.1870758. Delle Fave, A. and Massimini, F. (2005). The investigation of optimal experience and apathy: developmental and psychosocial implications. European Psychologist, 10, 264–74. Edwards, P. and Ramirez, P. (2016). When should workers embrace or resist new technology? New Technology, Work and Employment, 31(2), 99–113. El Zaatari, S., Marei, M., Li, W., and Usman, Z. (2019). Cobot programming for collaborative industrial tasks: an overview. Robotics and Autonomous Systems, 116, 162–80. Esteva, A., Kuprel, B., Novoa, R. A., Ko, J., Swetter, S. M., Blau, H. M., and Thrun, S. (2017). Dermatologist level classification of skin cancer with deep neural networks. Nature, 542(7639), 115–18. https://​doi​.org/​10​.1038/​nature21056. Fleming, P. (2019). Robots and organization studies: why robots might not want to steal your job. Organization Studies, 40(1), 23–38. https://​doi​.org/​10​.1177/​0170840618765568. Fullagar, C. and Delle Fave, A. (eds) (2017). Flow at Work: Measurement and Implications. New York: Taylor & Francis. https://​doi​.org/​10​.4324/​9781315871585​-2. Gasparre, A. and Tirabeni, L. (2020). Humans, robots and a typical day at work. A case-based research proposal. Conference: WOA 2020, ‘Will Employees Dream of The Electric Sheep? Impact of Digital Technologies Within and Beyond The Workplace’. Gibson, I., Rosen, D., Stucker, B., and Khorasani, M. (2021). Additive Manufacturing Technologies (3rd ed.). New York: Springer. https://​doi​.org/​10​.1007/​978​-3​-030​-56127​-7. Hackman, J. R. and Oldham, G. R. (1976). Motivation through the design of work: test of a theory. Organizational Behavior and Human Performance, 16(2), 250–79. Hackman, J. R. and Oldham, G. R. (1980). Work Redesign. Reading, MA: Addison-Wesley. Hoffman, G. (2019). Evaluating fluency in human–robot collaboration. IEEE Transactions on Human-Machine Systems, 49(3), 209–18. Huws, U. (2014). Labor in the Global Digital Economy. New York: Monthly Review Press. Jaehrling, K., Johnson, M., Larsen, T. P., Refslund, B., and Grimshaw, D. (2018). Tackling precarious work in public supply chains: a comparison of local government procurement policies in Denmark, Germany and the UK. Work, Employment and Society, 32 (3), 546–63. Jerčić, P., Hagelbäck, J., and Lindley, C. (2019). An affective serious game for collaboration between humans and robots. Entertainment Computing, 32, 100319. Kadir, B. A., Broberg, O., and Souza Da Conceição, C. (2018). Designing human-robot collaborations in industry 4.0: explorative case studies. Proceedings of International Design Conference, DESIGN, 2, 601–10. https://​doi​.org/​10​.21278/​idc​.2018​.0319. Kim, W., Lorenzini, M., Balatti, P., Nguyen, D. H. P., Pattacini, U., Tikhanoff, V., and Ajoudani, A. (2019). Adaptable workstations for human–robot collaboration: a reconfigurable framework for improving worker ergonomics and productivity. IEEE Robotics & Automation Magazine. Lanzing, M. (2016). The transparent self. Ethics and Information Technology, 18(1), 9–16. https://​doi​.org/​10​.1007/​s10676​-016​-9396​-y. Lazazzara, A., Tims, M., and de Gennaro, D. (2020). The process of reinventing a job: a meta-synthesis of qualitative job crafting research. Journal of Vocational Behavior, 116. https://​doi​.org/​10​.1016/​j​.jvb​.2019​.01​.001. Matheson, E., Minto, R., Zampieri, E. G. G., Faccio, M., and Rosati, G. (2019). Human-robot collaboration in manufacturing applications. A Review. Robotics, 8(4), 1–25. https://​doi​.org/​ 10​.3390/​robotics8040100.

268  Research handbook on human resource management and disruptive technologies

Moore, P. and Robinson, A. (2016). The quantified self: what counts in the neoliberal workplace. New Media & Society, 18(11), 2774–92. Morgeson, F. P. and Humphrey, S. E. (2006). The Work Design Questionnaire (WDQ): developing and validating a comprehensive measure for assessing job design and the nature of work. Journal of Applied Psychology, 91, 1321–39. Morgeson, F. P. and Humphrey, S. E. (2008). Job and team design: toward a more integrative conceptualization of work design. In J. Martoccio (ed.), Research in Personnel and Human Resource Management, 39–91. Bingley: Emerald Group Publishing Limited. http://​dx​.doi​ .org/​10​.1016/​S0742​-7301(08)27002​-7. Mühlemeyer, C. (2020). Assessment and design of employees-cobot-interaction. Advances in Intelligent Systems and Computing, 771–6. https://​doi​.org/​10​.1007/​978​-3​-030​-25629​ -6​_120. Neufeind, M., O’Reilly, J., and Ranft, F. (eds) (2018). Work in The Digital Age. London: Rowman & Littlefield International. Parker, S. K. (2014). Beyond motivation: job and work design for development, health, ambidexterity, and more. Annual Review of Psychology, 65(1), 661–91. Parker, S. K. and Grote, G. (2020). Automation, algorithms, and beyond: why work design matters more than ever in a digital world. Applied Psychology, 0, 1–45. Parker, S. K., Morgeson, F. P., and Johns, G. (2017a). One hundred years of work design research: looking back and looking forward. Journal of Applied Psychology, 102(3), 403. Parker, S. K., Van der Browck, A., and Holman, D. (2017b). Work design influences: a synthesis of multilevel factors that affect the design of jobs. The Academy of Management Annals, 11(1), 267–308. Parker, S. K., Wall, T. D., and Cordery, J. L. (2001). Future work design research and practice: towards an elaborated model of work design. Journal of Occupational and Organizational Psychology, 74, 413–40. Pham, Q., Madhavan, R., Righetti, L., Smart, W., and Chatila, R. (2018). The impact of robotics and automation on working conditions and employment [ethical, legal, and societal issues]. IEEE Robotics & Automation Magazine, 25, 126–8. Pollak, A., Paliga, M., Pulopulos, M. M., Kozusznik, B., and Kozusznik, M. W. (2020). Stress in manual and autonomous modes of collaboration with a cobot. Computers in Human Behavior, https://​doi​.org/​10​.1016/​j​.chb​.2020​.106469. Riek, L. D. (2017). Healthcare robotics. Communications of the ACM, 60(11), 68–78. Sanders, A., Elangeswaran, C., and Wulfsberg, J. (2016). Industry 4.0 implies lean manufacturing: research activities in industry 4.0 function as enablers for lean manufacturing. Journal of Industrial Engineering and Management, 9(3), 811–33. Schlogl, L., Weiss, E., and Prainsack, B. (2021). Constructing the ‘future of work’: an analysis of the policy discourse. New Technology, Work and Employment, 36(3), 307–26. Schumpeter, J.A. (2015). Digital Taylorism. The Economist. 12 September. Selby, A. (2017). Undercover at Amazon: Exhausted humans are inefficient, so robots are taking over. Mirror Online. https://​www​.mirror​.co​.uk/​news/​uk​-news/​undercover​-amazon​ -exhaustedhumans​-inefficient​-11593145. Senders, J. T., Arnaout, O., Karhade, A. V., Dasenbrock, H. H., Gormley, W. B., Broekman, M. L., Sherwani, K. I., Kumar, N., Chemori, A., Khan, M., and Mohammed, S. (2019). Rise-based adaptive control for exoskeleton to assist knee joint mobility. Robotics and Autonomous Systems, 124, 103354. Sherwani, F., Asad, M. M., and Ibrahim, B. S. K. K. (2020). Collaborative Robots and Industrial Revolution 4.0. Conference: 2020 International Conference on Emerging Trends in Smart Technologies (ICETST). Smids, J., Nyholm, S., and Berkers, H. (2019). Robots in the workplace: a threat to—or opportunity for—meaningful work? Philosophy and Technology, 33, 503–22. https://​doi​.org/​10​ .1007/​s13347​-019​-00377​-4.

Empowering or taking over?  269

Spencer, D. A. (2018). Fear and hope in an age of mass automation: debating the future of work. New Technology, Work and Employment, 33(1), 1–12. Srnicek, N. and Williams, A. (2015). Inventing the Future: Post Capitalism and a World Without Work. London and New York: Verso. van Wynsberghe, A., and Comes, T. (2020). Drones in humanitarian contexts, robot ethics, and the human–robot interaction. Ethics and Information Technology, 22(1), 43–53. https://​doi​ .org/​10​.1007/​s10676​-019​-09514​-1. Vanzo, A., Riccio, F., Sharf, M., Mirabella, V., Catarci, T., and Nardi, D. (2019). Who is willing to help robots? A user study on collaboration attitude. International Journal of Social Robotics, 1–10. Vasic, M. and Billard, A. (2013). Safety issues in human-robot interactions. IEEE International Conference on Robotics and Automation, 197–204. Villani, V., Pini, F., Leali, F., and Secchi, C. (2018). Survey on human–robot collaboration in industrial settings: safety, intuitive interfaces and applications. Mechatronics, 55, 248–66. Wang, B., Liu, Y., Qian, J., and Parker, S. K. (2020). Achieving effective remote working during the COVID-19 pandemic: a work design perspective. Applied Psychology, 70(1):16–59. Welfare, K. S., Hallowell, M. R., Shah, J. A., and Riek, L. D. (2019). Consider the human work experience when integrating robotics in the workplace. 14th ACM/IEEE International Conference on Human-Robot Interaction (HRI), 75–84. Wingfield, N. (2017). As Amazon pushes forward with robots, workers find new roles. The New York Times. https://​www​.nytimes​.com/​2017/​09/​10/​technology/​amazon​-robots​ -workers​.html.

19. The accelerating disconnection of work from time and place: new questions for HR Johannes Gartner, Kristiina Mäkelä, Jennie Sumelius and Hertta Vuorenmaa

INTRODUCTION Industrial revolutions have always been associated with major shifts in how people work, moving from agricultural and craft-based occupations to factory work in the first one, to assembly line work and scientific management in the second, and to knowledge work in the third industrial revolution (Bower and Christensen, 1995; Bodrožić and Adler, 2017). Currently, societies and organizations everywhere are experiencing a fourth industrial revolution, the key characteristic of which is an increasing blurring of the boundaries between the physical and the digital worlds (Schwab, 2017; Barley et al., 2017). A number of technologies have been associated with this widespread merger of the physical and the digital, ranging from artificial intelligence and robotics to the internet of things (IoT) and 3D printing, resulting in major changes in business models and production methods across industries. In addition to these industry-level changes, the fourth industrial revolution is also fundamentally changing the way in which people – white-collar knowledge workers, in particular – do their work. The simultaneous development of mobile, virtual and cloud technologies, and the network and broadband infrastructures that enable them, have disrupted the way in which work is carried out by making it possible to do more and more aspects of knowledge work from any location (not just the corporate office that has been the primary workplace in past decades) and at any time (not just during the traditional working hours of nine-to-five, Monday-to-Friday). Easier access to people and information means that knowledge work can be done in many different places and at flexible times, depending on either personal needs or organizational requirements: from homes and second homes, coworking spaces and coffee shops, and while travelling. These changes in ways of working will also have significant consequences for how human resources (HR) are managed going forward. Given that traditional models of human resource management (HRM) have been developed in relatively stable environments, in which most work takes place in corporate offices during office hours, they may prove less useful in the future. We, therefore, need to shift our focus from the effectiveness of our current practices towards understanding what the new timeand place-independent ways of working (Cooke et al., 2022) mean for HR. Against 270

The accelerating disconnection of work from time and place  271

this background, in this chapter we explore the HR implications of the disconnection of work from time and place, highlighting important new questions and concerns that HR needs to consider in the key areas of performance management, recruitment and talent management, training and development and diversity management. Rather than discussing the large number of emerging technologies that have been introduced during the last decade and their complex dependencies and interlinkages (Iansiti, 1995), this chapter focuses on how mobile, virtual and cloud technologies disrupt the way in which work is carried out, and through that, the work of the HR function. In what follows, we first discuss how work has evolved over time in relation to time and place, and then discuss how the current increasing disconnection might impact key HRM practices. We follow Ulrich and Dulebohn’s (2015) call for an increasingly externally oriented perspective on HRM and seek to widen the perspective of existing research on HRM and technology, which has largely focused on ‘creating value within and across organizations for targeted employees and management’ from an effectiveness perspective (Bondarouk and Ruël, 2009, 507). Our focus is specifically on white-collar knowledge work, referring to the application of domain-specific expertise to complex and novel problems (Van Der Vegt et al., 2006), as knowledge-workers are a business-critical target group of many central HRM processes and practices.

DISCONNECTION OF WORK FROM TIME AND PLACE Since the first and second industrial revolutions, salaried work has been closely linked to time. Taylorism and Fordism paced work by the clock (Baxter and Kroll-Smith, 2005). A 40-hour workweek was introduced in the first half of the twentieth century, with Monday-to-Friday nine-to-five becoming the standard office hours (Armstrong-Stassen, 1998; Bittman, 2016). With the evolution of personal computing and electronic communication tools such as e-mail, knowledge work started to expand past the traditional working week; the new tools enabled asynchronous work but also resulted in increasing perceptions of time pressure and spill-overs to leisure time (Godbey and Robinson, 1997; Wajcman, 2014; Bittman, 2016). This relatively bound relationship between work and time is now rapidly dissolving, as mobile, virtual and cloud technologies provide constantly improving access to information and people independently of time. As a consequence, the boundaries of work and non-work time are becoming increasingly blurred (Khallash and Kruse, 2012; Colbert et al., 2016). On the one hand, technology allows for a more blended, multipurpose timetable in which the individual has more freedom to choose and alternate between work and leisure time (Bittman, 2016; Colbert et al., 2016). Increased time flexibility reduces or avoids commuting time, and facilitates the management of personal matters during the day (Ryan and Wessel, 2015; Colbert et al., 2016). Although earlier research depicted the latter as a form of employee misbehaviour (Lim, 2002), more recent research has demonstrated that time autonomy facilitates employee wellbeing (Kattenbach et al., 2010; König and de la Guardia, 2014). On the

272  Research handbook on human resource management and disruptive technologies

other hand, work during the nonstandard evening or weekend hours has significantly increased (Eurofound, 2017a, 2017b), and the growing norm of constant availability (Mazmanian et al., 2013) has led to new concerns, such as collaborative overload (Cross et al., 2016), work-family conflict (Mäkelä et al., 2015), stress and exhaustion (Butts et al., 2015). The disconnection of work from time is intimately linked with the disconnection of work from a physical place. Issues of teleworking, telecommuting or distance-working have been examined from various angles since the 1980s (Handy, 1984; Hamblin, 1995; Valenduc and Vendramin, 2001; Golden and Raghuram, 2010), with the perceived fairness of different arrangements (Ryan and Wessel, 2015), identity building (Valenduc and Vendramin, 2016) and social ties receiving attention (Grantham, 2000). More recently, the geographies of knowledge and collaboration have become increasingly manifold (Ellem, 2016). Most knowledge-intensive tasks can be performed on our own laptops thanks to increased computing power and cloud access, and virtual and mobile collaboration technologies enable work across distance, not just from home but anywhere (Alghamdi et al., 2016; Standaert et al., 2016). Talent and expertise are more dispersed geographically, and working in global virtual teams has fast become the modus operandi of multinational organizations in particular (Zander et al., 2012; Nurmi and Hinds, 2016). The pandemic has further accelerated remote working arrangements, leading to record levels of mobility in the job market (OECD, 2021). In sum, the disconnection of work from time and place has been developing over a long time, with the COVID-19 pandemic accelerating and intensifying the development significantly (Grömling, 2021). In terms of place, remote and hybrid work (the latter combining office-based work with working from home or third places) are likely to become permanent features of future working life. Relatedly, the blurring and intertwining of work- and life-time will also likely expand, both in terms of life seeping into traditional work times and work spilling over to all times (Teodorovicz et al., 2021). These changes bring many new questions for employers and their human resource management practices, to which we turn next.

NEW QUESTIONS RELATED TO CURRENT HRM PRACTICES Above, we have described the accelerating process of technology-driven change in white-collar knowledge work, using the two interlinked disconnections of work from time and place as organizing labels. We now move on to consider what these disconnections mean for human resource management. We first discuss potential new questions related to the current HRM practices of performance management, recruitment and talent management, training and development and diversity management, and then move on to exploring new elements that HR will need to focus on going forward. These new elements have to do with building employee engagement and an inspiring organizational culture in a virtual and hybrid environment, focusing on the

The accelerating disconnection of work from time and place  273

sustainability of working life and proactive employee wellbeing to a much greater extent than before, and facilitating a resilient mindset among employees that allows them to thrive in a constantly changing environment. New Questions Related to Performance Management Performance management practices have already, for some time, been criticized for being overly static, heavy and process-driven (Cappelli and Tavis, 2016), which is at odds with the rapidly evolving business environment. This has, in parallel with the changing role of time and place, increased interest among both practitioners and researchers for technology-aided, real-time performance management practices. Performance management is moving away from time (annual cycle of once- or twice-yearly meetings) and place (relatively standardized face-to-face supervisor meetings in the office)-bound processes towards more output- and results-based approaches characterized by real-time data, dashboards and pulse surveys (Curzi et al., 2019). This real-time and data-driven approach requires more from HR, in that goals need to be clear and cascade down the organization, data has to be open and transparent and systems need to be digital and accessible. This requires significant investments in digital systems, data capabilities and process design, and many organizations are already moving in these directions. A more challenging development is the recent emergence of electronic performance monitoring (EPM) capabilities that enable digital observation of measures such as keyboard strokes, mouse movements and email use (Jeske, 2021) or log-in/ log-off times and working pace (Kalischko and Riedl, 2021). On the positive side, EPM tools enable data-driven management (Schwarzmüller et al., 2018), which can add value by providing insight into effective ways of working. On the other side, employees dislike continuous monitoring (Jeske and Santuzzi, 2015), which is perceived as dehumanizing and stressful (Jeske, 2021) and can lead to mistrust and lower job satisfaction (Kalischko and Riedl, 2021). The use of EPMs also relates to important code of conduct considerations that HR needs to pay attention to going forward. A central question is how EPMs are used in an ethical way (Ravid et al., 2020); for example, the implementation of EPMs needs to be coupled with transparency to employees about what exactly is being monitored, at which level of anonymity, when, where, and why. Information privacy and data protection concerning individuals (Bélanger and Crossler, 2011; Carpenter et al., 2016) is a core issue and is currently approached quite differently depending on national cultures, laws and regulations (Kalischko and Riedl, 2021). The jury is still out on these technologies, but given their rapid development, HR will likely need to take an operational and ethical stand on them sooner rather than later. New Questions Related to Recruitment and Talent Management The disconnection of work from time and place opens up global talent markets more broadly than before – an increasing number of jobs in an increasing number of fields

274  Research handbook on human resource management and disruptive technologies

can be done from anywhere at any time. This means that the employee value proposition to attract the best talent becomes more important than ever before. Organizations and their HR have to ask themselves why the increasingly mobile talent, the best of whom are in high demand, would want to work for them. These ‘stars’ will have more options than ever through remote and hybrid work, new online platforms, networks and digital marketplaces. Amongst the immediate issues facing HR are the commitment and retention related questions (e.g. Morris et al., 2016), which have accelerated during the COVID-19-induced ‘great resignation’ (Sull et al., 2022). As a response, HR may need to find more tailored and personal-needs-based employment contracts for talent that include flexible remote and hybrid arrangements, customized work content, personalized development opportunities or work-life integration (Scholarios and Marks, 2004). We also see a related shift away from traditional hierarchical career ladders to more individualistic portfolios of meaningful projects, coupled with increasing mobility across organizational and also professional or occupational boundaries. If there is a move towards measuring success not by within-company hierarchical advancement, but by the achievement of personal goals, self-development and psychologically meaningful work portfolios (Banai and Harry, 2004; Sullivan and Baruch, 2009), this means shifting competition for recruitment from more status-driven positions to more values-driven projects. Relatedly, although the benefits of long-term work arrangements may prevail, an important question is whether these more personally driven work arrangements will move from the employment domain to that of entrepreneurial and/or contractual work (Capelli and Keller, 2013). If the talents increasingly work with multiple simultaneous assignments and employers, how do we ensure their commitment to our project? Less skilled employees, for their part, are more likely to be left outside traditional permanent employment: either contracted temporarily as needed with potentially fluctuating wages, removed altogether as a result of automatization and robotization, or moved to low-cost areas on a much broader basis than today, made possible by telepresence and augmented reality (Kristoffersson et al., 2013). New Questions Related to Training and Development Continuous, holistically designed, lifelong training is positively and directly related to organizational performance (Garavan et al., 2021a, 2021b), and increasingly a necessity in all areas of work. Technological development makes extant skills rapidly obsolete, which has put continuous education firmly on the agenda for most governments (Eurofound, 2017a, 2017b), both in terms of securing jobs in the country and ensuring that there are enough skilled employees for those jobs. For HR, the key question becomes whether companies should be responsible for lifelong learning or if the responsibility resides primarily with the individual. And who should fund the training – companies, individuals or governments (Hytti and O’Gorman, 2004; Kost et al., 2020). The answer is probably all of these, but solutions for this question are currently being debated and are yet to be worked out.

The accelerating disconnection of work from time and place  275

Organizations have responded to the growing pressures by raising upskilling high on the strategic agenda and investing in training and development, even if COVID-19 has created a slight drop in these figures (Kieran et al., 2021). There is research evidence that the strength of the relationship between training and organizational performance has increased every year over time, as the complexity of the business environment has increased (Garavan et al., 2021a, 2021b). There is, however, less clarity on what types of skills will be most important for white-collar employees in the future. Some stress hard skills such as technological knowledge, problem-solving and logical thinking (Tuomi et al., 2018), while others predict that soft skills such as creativity and emotional intelligence will hold increasing significance in coming decades (Gray, 2016). A third group emphasizes digital skills such as IT competencies, teleworking abilities and digital communication skills as enablers (Schwarzmüller et al., 2018). This breadth of needed skills means that HR needs to adopt a broad, holistic perspective when analysing and planning organizational training and development needs and investments, including both general and firm-specific training (Garavan et al., 2021a, 2021b). Garavan et al. (2021a) further suggest that HR should be able to do an organizational skills analysis and have the skills levels and needs of their own employees as their starting point when designing training and development. New Questions Related to Diversity Management Diversity management, or promoting the inclusion of employees from different backgrounds in the organization, has been an HR focus area for some time already. While previous literature has acknowledged the importance and potential benefits of a diverse workforce, it has simultaneously recognized that simply having a diverse workforce does not guarantee the realization of benefits for the organization or the individual (Randel et al., 2018). The COVID-19 pandemic has deepened existing inequalities in working life, creating a further imperative for focusing on diversity and inclusion in the workforce (ILO, 2021; Lee et al., 2022). What is more, as particularly top talents’ employment arrangements are likely to become more personalized, they need to be coupled with growing transparency. The complexities of managing an increasingly diverse workforce will require a more holistic approach from HR than before (Triana et al., 2021), broadening the focus of diversity management from equal opportunities towards inclusion. Inclusive organizations not only acknowledge the importance and potential benefits of diverse backgrounds among personnel (Randel et al., 2018), but also strive to adopt practices that enable the full participation and contribution of everyone (Roberson, 2006; Ferdman, 2014). A key question for HR thus becomes how to ensure that all employees have the necessary skills and opportunities to fully participate in and contribute to the organization despite their background, where they work from and whether they work in a face-to-face or virtual mode. This includes aspects such as recruiting individuals with an inclusive and open mindset (Richard et al., 2013; Noon and Ogbonna, 2021), reviewing existing HRM practices and ways of working from an inclusion and

276  Research handbook on human resource management and disruptive technologies

identity-conscious perspective (Roberson et al., 2020) and fostering an environment open to different points of view (Ragins and Ehrhardt, 2020).

DISCUSSION As we have outlined above, the nature of white-collar work is changing rapidly as a result of technological developments, and the COVID-19 pandemic has further accelerated the change. New ways of working and managing people are emerging; some are discussed in this chapter, while others are not yet foreseeable. In this chapter we used two disconnections as a way of conceptualizing some of the key changes – the disconnection between work and time, and the disconnection between work and place – and discussed the implication of these changes for HR in terms of how existing HRM practices may need to be adapted and redesigned. The changes also mean that, in addition to rethinking existing HRM practices, HR needs to expand its role to address new elements that have not been considered in earlier research or practice to a sufficient extent, or much at all. When work is increasingly disconnected from time and place, organizing it becomes more complex and heterogeneous than before, and individuals need to be able to function in a more uncertain and ambiguous environment. Addressing this increasing complexity and uncertainty with an overall aim of finding more socially sustainable ways of managing people in the hybrid workplace requires new capabilities from HR, both in terms of managing tensions, and recognizing the challenges of aligning multiple perspectives with reduced resources (Collings et al., 2021). First, HR needs to be able to build employee engagement and an inspiring organizational culture in a working environment that is characterized by remote, virtual and hybrid working modes. Remote work increases autonomy and flexibility in that the individual has more degrees of freedom to choose how to work, it can also negatively influence the physical, emotional and social working environments of employees, and it is often associated with perceived isolation from social aspects of organizations (Wiesenfeld et al., 2001). Previous research has shown that lack of embeddedness in an organization carries risks in terms of decreased wellbeing and increased employee turnover (Mitchell et al., 2001). Due to the pandemic, this risk is now more actual than ever, as widespread remote work and separation has led to an erosion of engagement and shared organizational culture (Sull et al., 2022). Second, the sustainability of working life in the constantly changing environment requires HR to shift emphasis from occupational healthcare interventions to promoting employee wellbeing more proactively. Blurring work-life boundaries allows individuals to organize their work to better suit their individual needs, but it also creates pressure to be constantly available and reachable. The mental burden of balancing job demands and resources (Bakker and Demerouti, 2007), and constant connectivity and interruptions in particular, have emerged as central concerns over and above the previously dominant physical safety issues (Eurofound, 2017b).

The accelerating disconnection of work from time and place  277

CONCLUSION AND RECOMMENDATIONS FOR ACTION As work will continue to disconnect from time and place, a key area in which HR needs to take action concerns how to build employee engagement and organizational culture in remote and virtual working modes. The mindset of the HR function and management will also need to be reconsidered. HR research and practice has traditionally focused on the question of whether organizational units such as subsidiaries of a multinational corporation should have globally standardized versus locally responsive HRM policies, practices and service delivery (Rosenzweig and Nohria, 1994). In light of the ongoing changes, this may no longer be the most important distinction. Instead, HR may need to think increasingly about individualized and flexible work arrangements. This leads to the key HRM question being not between global standardization and local responsiveness, but rather between standardization and individualization. How do firms design practices that facilitate all the change and can be tailored to individuals rather than groups or locations? Such an HR-as-a-facilitator model may become a mass-adapted modular and stackable portfolio of different services that individuals can have access to, or choose from, and forgo the current global versus local dilemma altogether. In a similar vein, organizations may need to gradually move away from line-management supervision towards a more trust-based self-management model, where employees have higher personal responsibility and self-leadership. HR needs to act as a facilitator of the required skills and modes of working. Furthermore, previously nonexisting forms of HR, such as digital marketplace HRM, may altogether replace some of the activities currently executed by company-internal HR. The increased blurriness of work-life boundaries that comes with technological developments means that an urgent new question for HR is how to develop policies and practices to protect and look after employee wellbeing in a proactive way, before occupational healthcare interventions are needed (Cleveland et al., 2015). Wellbeing literature has thus far developed as a largely separate stream from HR literature (Guest, 2017), but the time has come to connect research findings and practical learning from both. The new challenges concerning existing HRM practices also require new capabilities and competencies from HR professionals. In addition to, or in lieu of, the more traditional educational backgrounds of (organizational) psychology, business and management and (adult) education, HR professionals will need to have increasingly strong data and statistical skills. HR managers need to master new technologies, which will also likely decrease the need for traditional HR staff – but these new technologies will potentially also free up time for HR professionals to focus on more strategic issues. Another issue concerns the skills required from HR professionals in managing the increasing virtuality of the workplace. Dealing with people and culture-related issues virtually, including complicated and emotionally heavy issues such as downsizing, is a major challenge. Individual-level issues such as under-performance or illness will also be further complicated by distance and virtual communication. Lastly, the importance of diversity and inclusion and other

278  Research handbook on human resource management and disruptive technologies

ethical questions is rising, and researchers have expressed concerns that current HR education might no longer be sufficient when dealing with the growing complexities of managing a diverse workforce (Triana et al., 2021). There are also other more macro-level changes driven by technological development, and by automation and robotization in particular, that will require dedicated analysis to be properly unpacked. The implications of all these changes – the disappearance of many types of jobs and the resulting creation of a ‘new precariat’ (Standing, 2014) – go well beyond HR. Although they have far-reaching consequences on all levels of analysis from psychological to societal and political, they are outside the scope of this paper. Ethical questions alone pose a serious professional challenge, as current legal frameworks largely lag behind technological developments. Although challenging, the increased need for ethical considerations may also offer a new way for HR to assert its relevance. Finally, the employee-organization relationship is also changing at an accelerating pace (Coyle-Shapiro and Shore, 2007; Ryan and Wessel, 2015), with employees no longer being bound to their organizations in the same way as before. Emerging electronic platforms and network-based business models (e.g. Stanford, 2017) will provide viable opportunities for more types of workers and professionals than today to sell their expertise through self-managed portfolios both locally and globally. For HR, this raises new questions related to things like how to attract and retain individual freelancers/entrepreneurs, how to compete for them internationally, how to engage them and how to create contracts that hinder them from spilling confidential elements of their assignments. Recruitment may need to be approached differently – what is on offer, if not employment? On the other hand, matching sellers and buyers of work can be fully automated with the help of various platforms, crowd feedback and artificial intelligence, removing many HR tasks altogether.

REFERENCES Alghamdi, M., Regenbrecht, H., Hoermann, S., Langlotz, T., and Aldridge, C. (2016). Social presence and mode of videocommunication in a collaborative virtual environment. In PACIS (p. 126), June. Armstrong-Stassen, M. (1998). Alternative work arrangements: meeting the challenges. Canadian Psychology/Psychologie Canadienne, 39(1–2), 108. Bakker, A. B. and Demerouti, E. (2007). The job demands-resources model: state of the art. Journal of Managerial Psychology, 22(3), 309–28. Banai, M. and Harry, W. (2004). Boundaryless global careers: the international itinerants. International Studies of Management & Organization, 34(3), 96–120. Barley, S. R., Bechky, B. A., and Milliken, F. (2017). From the editor: the changing nature of work: careers, identities, and work lives in the 21st century. Academy of Management Discoveries, 3(2), 111–15. Baxter, V. and Kroll-Smith, S. (2005). Normalizing the workplace nap: blurring the boundaries between public and private space and time. Current Sociology, 53(1), 33–55. Bélanger, F. and Crossler R. E. (2011). Privacy in the digital age: a review of information privacy research in information systems. MIS Q, 35, 1017–41. 10.2307/41409971.

The accelerating disconnection of work from time and place  279

Bittman, M. (2016). Working time. In S. Edgell, G. Gottfried, and E. Granter (eds), Handbook of the Sociology of Work and Employment. London: Sage. Bodrožić, Z. and Adler, P. S. (2017). The evolution of management models: a neo-Schumpeterian theory. Administrative Science Quarterly, 0001839217704811. Bondarouk, T. V. and Ruël, H. J. M. (2009). Electronic human resource management: challenges in the digital era. International Journal of Human Resource Management, 20(3), 505–14. Bower, J. L. and Christensen, C. M. (1995). Disruptive technologies: catching the wave. Harvard Business Review, January–February. Butts, M. M., Becker, W. J., and Boswell, W. R. (2015). Hot buttons and time sinks: the effects of electronic communication during nonwork time on emotions and work-nonwork conflict. Academy of Management Journal, 58(3), 763–88. Cappelli, P. and Keller, J. R. (2013). Classifying work in the new economy. Academy of Management Review, 38(4), 575–96. Cappelli, P. and Tavis, A. (2016). The performance management revolution. Harvard Business Review, 94(10), 58–67. Carpenter, D., McLeod, A., and Hicks C. (2016). Privacy and biometrics: an empirical examination of employee concerns. Information Systems Frontiers, 20, 91–110. 10.1007/ s10796-016-9667-5. Cleveland, J. N., Byrne, Z. S., and Cavanagh, T. M. (2015). The future of HR is RH: respect for humanity at work. Human Resource Management Review, 25(2), 146–61. Colbert, A., Yee, N., and George, G. (2016). The digital workforce and the workplace of the future. Academy of Management Journal, 59(3), 731–9. Collings, D. G., McMackin, J., Nyberg, A. J., and Wright, P. M. (2021). Strategic human resource management and COVID-19: emerging challenges and research opportunities. Journal of Management Studies, 58(5), 1378–82. Cooke, F. L., Dickmann, M., and Parry, E. (2022). Building sustainable societies through human-centred human resource management: emerging issues and research opportunities. The International Journal of Human Resource Management, 33(1), 1–15. Coyle-Shapiro, J. A. and Shore, L. M. (2007). The employee–organization relationship: where do we go from here?. Human Resource Management Review, 17(2), 166–79. Cross, R., Rebele, R., and Grant, A. (2016). Collaborative overload. Harvard Business Review, January–February: 75–9. Curzi, Y., Fabbri, T., Scapolan, A. C., and Boscolo, S. (2019). Performance appraisal and innovative behavior in the digital era. Frontiers in Psychology, 10, 1659. Ellem, B. (2016). Geographies of the labour process: automation and the spatiality of mining. Work, Employment and Society, 30(6), 932–48. Eurofound (2017a). Aspects of Non-Standard Employment in Europe. Luxembourg: Publications Office of the European Union. Eurofound (2017b). Sixth European Working Conditions Survey – Overview Report (2017 update). Luxembourg: Publications Office of the European Union. Ferdman, B. M. (2014). The practice of inclusion in diverse organizations: toward a systemic and inclusive framework. In B. M. Ferdman and B. R. Deane (eds), Diversity at Work: The Practice of Inclusion (pp. 3–54). Jossey-Bass/Wiley. Garavan, T. N., McCarthy, A., Lai, Y., Clarke, N., Carbery, R., Gubbins, C., ... and Saunders, M. N. (2021a). Putting the system back into training and firm performance research: a review and research agenda. Human Resource Management Journal, 31(4), 870–903. Garavan, T., McCarthy, A., Lai, Y., Murphy, K., Sheehan, M., and Carbery, R. (2021b). Training and organisational performance: a meta analysis of temporal, institutional and organisational context moderators. Human Resource Management Journal, 31(1), 93–119. Godbey, G. and Robinson, J. (1997). The increasing prospects for leisure. Parks & Recreation (Arlington), 32(6), 74–82.

280  Research handbook on human resource management and disruptive technologies

Golden, T. D. and Raghuram, S. (2010). Teleworker knowledge sharing and the role of altered relational and technological interactions. Journal of Organizational Behavior, 31(8), 1061–85. Grantham, C. (2000). The Future of Work: The Promise of the New Digital Work Society. New York: McGraw-Hill. Gray, H. (2016). Precarious diversity: representation and demography. In M. Curtin and K. Sanson (eds), Precarious Creativity (pp. 241–53). California: University of California Press. Grömling, M. (2021). COVID-19 and the growth potential. Intereconomics, 56(1), 45–9. Guest, D. E. (2017). Human resource management and employee well-being: towards a new analytic framework. Human Resource Management Journal, 27(1), 22–38. Hamblin, H. (1995). Employees’ perspectives on one dimension of labour flexibility: working at a distance. Work, Employment and Society, 9(3), 473–98. Handy, C. (1984). The Future of Work: What Jobs Will There Be? What Will Life Be Like? What Needs to Be Done? New York: Basil Blackwell. Hytti, U. and O’Gorman, C. (2004). What is ‘enterprise education’? An analysis of the objectives and methods of enterprise education programmes in four European countries. Education and Training, 46, 11–23. Iansiti, M. (1995). Technology integration: Managing technological evolution in a complex environment. Research Policy, 24(4), 521–42. ILO (2021). Research Brief: COVID-19 and rising wage inequality: trends and challenges in Thailand and Viet Nam. January 2021. https://​www​.ilo​.org/​wcmsp5/​groups/​public/​-​-​-asia/​ -​-​-ro​-bangkok/​documents/​briefingnote/​wcms​_767475​.pdf. Jeske, D. (2021). Monitoring remote employees: implications for HR. Strategic HR Review, 20(2), 42–6. Jeske, D. and Santuzzi, A. M. (2015). Monitoring what and how: psychological implications of electronic performance monitoring. New Technology, Work and Employment, 30(1), 62–78. Kalischko, T. and Riedl, R. (2021). Electronic performance monitoring in the digital workplace: conceptualization, review of effects and moderators, and future research opportunities. Frontiers in Psychology, 12. Kattenbach, R., Demerouti, E., and Nachreiner, F. (2010). Flexible working times: effects on employees’ exhaustion, work-nonwork conflict and job performance. Career Development International. Khallash, S. and Kruse, M. (2012). The future of work and work-life balance 2025. Futures, 44(7), 678–86. Kieran, S., MacMahon, J., and MacCurtain, S. (2021). Simple rules for sensemaking praxis: how HR can contribute to strategic change by developing sensemaking capability in organisations. Human Resource Management Journal, 32(4). König, C. J. and de la Guardia, M. E. C. (2014). Exploring the positive side of personal internet use at work: does it help in managing the border between work and nonwork?. Computers in Human Behavior, 30, 355–60. Kost, D., Fieseler, C., and Wong, S. I. (2020). Boundaryless careers in the gig economy: an oxymoron? Human Resource Management Journal, 30(1), 100–13. Kristoffersson, A., Coradeschi, S., and Loutfi, A. (2013). A review of mobile robotic telepresence. Advances in Human-Computer Interaction, 2013, 1–17. Lee, J. Y., Yahiaoui, D., Lee, K. P., and Cooke, F. L. (2022). Global talent management and multinational subsidiaries’ resilience in the Covid-19 crisis: moderating roles of regional headquarters’ support and headquarters–subsidiary friction. Human Resource Management, 61(1). Lim, V. K. G. (2002). The IT way of loafing on the job: cyberloafing, neutralizing and organizational justice. Journal of Organizational Behaviour, 23, 675–94.

The accelerating disconnection of work from time and place  281

Mäkelä, L., Kinnunen, U., and Suutari, V. (2015). Work to life conflict and enrichment among international business travelers: the role of international career orientation. Human Resource Management, 54(3), 517–31. Mazmanian, M., Orlikowski, W. J., and Yates, J. (2013). The autonomy paradox: the implications of mobile email devices for knowledge professionals. Organization Science, 24(5), 1337–57. Mitchell, T. R., Holtom, B. C., Lee, T. W., Sablynski, C. J., and Erez, M. (2001). Why people stay: using job embeddedness to predict voluntary turnover. Academy of Management Journal, 44(6), 1102–21. Morris, S., Snell, S., and Björkman, I. (2016). An architectural framework for global talent management. Journal of International Business Studies, 47(6), 723–47. Noon, M. and Ogbonna, E. (2021). Controlling management to deliver diversity and inclusion: prospects and limits. Human Resource Management Journal, 31(3), 619–38. Nurmi, N. and Hinds, P. J. (2016). Job complexity and learning opportunities: a silver lining in the design of global virtual work. Journal of International Business Studies, 47(6), 631–54. OECD (2021). OECD Employment Outlook 2021 – Navigating the COVID-19 Crisis and Recovery. 7 July. https://​doi​.org/​10​.1787/​5a700c4b​-en. Ragins, B. R. and Ehrhardt, K. (2020). Gaining perspective: the impact of close cross-race friendships on diversity training and education. Journal of Applied Psychology, 106(6), 856–81. https://​doi​.org/​10​.1037/​apl0000807. Randel, A. E., Galvin, B. M., Shore, L. M., Ehrhart, K. H., Chung, B. G., Dean, M. A., and Kedharnath, U. (2018). Inclusive leadership: realizing positive outcomes through belongingness and being valued for uniqueness. Human Resource Management Review, 28(2), 190–203. Ravid, D. M., Tomczak, D. L., White, J. C., and Behrend, T. S. (2020). EPM 20/20: a review, framework, and research agenda for electronic performance monitoring. Journal of Management, 46(1), 100–26. Richard, O. C., Roh, H., and Pieper, J. (2013). The impact of diversity practices on racial diversity in management: the moderating role of firm size. Human Resource Management, 52(2), 215–42. Roberson, Q. M. (2006). Disentangling the meanings of diversity and inclusion in organizations. Group & Organization Management, 31(2), 212–36. Roberson, Q., King, E., and Hebl, M. (2020). Designing more effective practices for reducing workplace inequality. Behavioral Science & Policy, 6(1), 39–49. Rosenzweig, P. M. and Nohria, N. (1994). Influences on human resource management practices in multinational corporations. Journal of International Business Studies, 25(2), 229–51. Ryan, A. M. and Wessel, J. L. (2015). Implications of a changing workforce and workplace for justice perceptions and expectations. Human Resource Management Review, 25(2), 162–75. Scholarios, D. and Marks, A. (2004). Work-life balance and the software worker. Human Resource Management Journal, 14(2), 54–74. Schwab, K. (2017). The Fourth Industrial Revolution. London: Currency, Illustrated Edition (3 January 2017) ISBN-10:9781524758868; ISBN-13:978-1524758868. Schwarzmüller, T., Brosi, P., Duman, D., and Welpe, I. M. (2018). How does the digital transformation affect organizations? Key themes of change in work design and leadership. Management Revue, 29(2), 114–38. Standaert, W., Muylle, S., and Basu, A. (2016). An empirical study of the effectiveness of telepresence as a business meeting mode. Information Technology and Management, 17(4), 323–39. Standing, G. (2014). The Precariat: The New Dangerous Class. New York: Bloomsbury Academic.

282  Research handbook on human resource management and disruptive technologies

Stanford, J. (2017). The resurgence of gig work: historical and theoretical perspectives. The Economic and Labour Relations Review, 28(3), 382–401. Sull, D., Sull, C., and Zweig, B. (2022). Toxic culture is driving the great resignation. MIT Sloan Management Review, 63(2), 1–9. Sullivan, S. E. and Baruch, Y. (2009). Advances in career theory and research: a critical review and agenda for future exploration. Journal of Management, 35(6), 1542–71. Teodorovicz, T., Sadun, R., Kun, A. L., and Shaer, O. (2021). How does working from home during COVID-19 affect what managers do? Evidence from time-use studies, Human– Computer Interaction. 10.1080/07370024.2021.1987908. Triana, M. D. C., Gu, P., Chapa, O., Richard, O., and Colella, A. (2021). Sixty years of discrimination and diversity research in human resource management: a review with suggestions for future research directions. Human Resource Management, 60(1), 145–204. Tuomi, P., Multisilta, J., Saarikoski, P., and Suominen, J. (2018). Coding skills as a success factor for a society. Education and Information Technologies, 23(1), 419–34. Ulrich, D. and Dulebohn, J. H. (2015). Are we there yet? What’s next for HR? Human Resource Management Review, 25(2), 188–204. Valenduc, G. and Vendramin, P. (2001). Telework: from distance working to new forms of flexible work organisation. Transfer: European Review of Labour and Research, 7(2), 244–57. Valenduc, G. and Vendramin, P. (2016). Work in the Digital Economy: Sorting the Old from the New (vol. 3). Brussels: European Trade Union Institute. Van Der Vegt, G. S., Bunderson, J. S., and Oosterhof, A. (2006). Expertness diversity and interpersonal helping in teams: why those who need the most help end up getting the least. Academy of Management Journal, 49(5), 877–93. Wajcman, J. (2014). Pressed for Time: The Acceleration of Life in Digital Capitalism. Chicago: University of Chicago Press. Wiesenfeld, B. M., Raghuram, S., and Garud, R. (2001). Organizational identification among virtual workers: the role of need for affiliation and perceived work-based social support. Journal of Management, 27(2), 213–29. Zander, L., Mockaitis, A. I., and Butler, C. L. (2012). Leading global teams. Journal of World Business, 47(4), 592–603.

20. Keep in touch in remote workplaces: the relationship between collegial isolation and contextual work performance in remote work settings and the mediating role of relatedness Pascale Peters, Robert Jan Blomme, Martine Coun and Max Weijers

INTRODUCTION During the COVID-19 pandemic, which was announced in March 2020, employees were forced to shift to substantial, oftentimes full-time, tele(home)working, referring to a remote way of working enabled by information and communication technology (ICT) (Anderson and Kelliher, 2020). The intensification of the tele(home)working practice, and the associated use of disruptive technologies, such as Skype, Microsoft Teams, Zoom, or alternative technologies that enable remote working at higher speed, with less latency and enhanced flexibility, was shown to have both advantages and disadvantages. A study by the Dutch Knowledge Institute for Mobility Policy (Hamersma et al., 2020) conducted during the beginning of the pandemic, for example, showed that tele(home)workers experienced greater time effectiveness and focus and, hence, better work performance, than before the pandemic. In line with this, tele(home)workers in the study by Van Veldhoven and Van Gelder (2020) indicated that because of the intensive tele(home)work practice, they had more control over their working day, were better able to work in line with their personal preferences and worked more efficiently. On the other hand, Yerkes et al. (2020) showed that the work-life interference associated with tele(home)working during the pandemic was particularly demanding for employees with children, but that also single and childless workers experienced negative consequences of tele(home)working. In fact, earlier telework research already suggested that these latter categories are most at risk of loneliness, a sense of aimlessness and, relatedly, to experience negative effects on their wellbeing (Achor et al., 2018). Also Van Veldhoven and Van Gelder (2020) showed that not being colocated with colleagues due to the pandemic affected employees’ perceptions of social relationships at work, which relates to the concept of professional isolation. Professional isolation, or alternatively, workplace isolation, refers to a psychological construct related to employees’ ‘perceptions of isolation from the organization and from co-workers. Isolation perceptions are formed by the absence of support 283

284  Research handbook on human resource management and disruptive technologies

from co-workers and supervisors and the lack of opportunities for social and emotional interactions with the team’ (Marshall et al., 2007, p. 198). It has been shown that the more people work from home, the more professional isolation they experience, even when ICT allows them to communicate with relevant others (Wiesenfeld et al., 2001; Marshall et al., 2007; Golden et al., 2008; Orhan et al., 2016; Golden and Gajendran, 2018), which can be attributed to the reduced possibilities for interactions and the communication with peers and supervisors being mediated by ICT (Kurland and Bailey, 1999; Cooper and Kurland, 2002). The lack of (direct) interactions can enhance feelings of missing out, for example, on information and career opportunities (Allen et al., 2015). In line with this, Marshall et al. (2007) distinguish two dimensions of professional isolation: organizational isolation, which refers to the fear of missing out on being part of a network within the organization and experiencing a lack of support from the organization, and collegial isolation, which refers to missing a bond with colleagues, having colleagues around, and talking informally with colleagues, which enables the individual to achieve work-related goals. The telework intensity and duration since the COVID-19 pandemic may have increased the risk of collegial isolation and, consequently, stress and a loss of productivity (Toscano and Zappalà, 2020). Golden et al. (2008) argued that lacking (physical) contact with colleagues and not having the ability to mirror and test one’s behavior can cause employees to have less confidence in their abilities and knowledge and, hence, their work performance. Individual work performance can be defined as: ‘behaviors or actions of employees that are relevant to the objective of the organization’ (Koopmans et al., 2014, p. 231). Three subdimensions can be distinguished: job-specific or task-performance, contextual performance and counterproductive work performance (Koopmans et al., 2014). Task performance entails competence in planning and organizing work, focus on the quality of the work delivered and the results and the ability to work efficiently. Contextual performance refers to employees’ behaviors and actions that are developed outside of the employee’s main tasks and that support the organization, the employee and others, such as taking on additional and challenging tasks, taking the initiative, developing knowledge and skills and motivating and coaching others. This type of behavior may be particularly important in disruptive situations, such as the COVID-19 pandemic, in which routines may not suffice anymore. Counterproductive performance includes various types of behaviors, such as complaining, taking actions that endanger the organization, misusing information, time and resources and delivering poor quality work, which has a negative value for the effectiveness of an organization. Particularly regarding the subdimension of contextual work performance, including activities such as keeping track of knowledge, taking on new challenges beyond one’s formal job description and actively participating in meetings (Orr et al., 1989), interaction and communication with others is needed (Borman and Motowidlo, 1997; Motowidlo et al., 1997). This, however, was strongly compromised during the COVID-19 pandemic, when employees were enforced to work from home and, consequently, were more isolated from colleagues.

Keep in touch in remote workplaces  285

Therefore, the question rises how in a work context in which disruptive technology is used, such as intensified and enduring (mandatory) homeworking, in which people can experience professional isolation, individuals’ (contextual) work performance is affected. More specifically, due to the enhanced tele(home)work intensity and duration due the COVID-19 measures, face-to-face interactions among colleagues could not take place, which might have inhibited employees in building up a (close) bond with their colleagues. Collegial isolation, in turn, may limit an employee’s feeling of being part of a larger group, which can reduce their autonomous motivation to engage in contextual work activities (cf. Galanti et al., 2021; Sardeshmukh et al., 2012). The explanation for this possible chain of events may be found in the self-determination theory (SDT) (Deci and Ryan, 2000), which assumes that everyone has certain basic psychological needs (i.e. autonomy, competence and relatedness) that can be satisfied by conditions in both the work and private domains. At work, particularly the basic need of relatedness may come under pressure when employees experience collegial isolation, for example, coming at the expense of their motivation and opportunity to invest in personal growth and development. In the light of the account above, the present cross-sectional study conducted during the COVID-19 pandemic in which employees had to work from home for a substantial part of their working hours aims to provide insight into the relationship between collegial isolation and individual contextual work performance and the mediating role of relatedness in this relationship. The contributions to the literature are twofold. First, the study contributes to the debate of teleworking by shedding light on the consequences of professional isolation, and more specifically collegial isolation, on contextual work performance, a form of individual performance that demands collegial interaction, and is increasingly important in disruptive work contexts in which employees have to proactively manage themselves and stimulate others, take responsibilities and find creative solutions for unexpected tasks, such as the crisis situations characterizing the COVID-pandemic and future hybrid work contexts. Second, the study contributes to the SDT-literature (Deci and Ryan, 2000; Van den Broeck et al., 2010) by examining the mediating role of relatedness in the relationship between collegial isolation (in the context of the COVID-19 pandemic) and individuals’ contextual work performance. The need for relatedness at work can be satisfied when employees experience being part of a meaningful social and caring work community. Feeling connected and maintaining personal relationships with others can lead to higher psychological wellbeing (self-esteem) and autonomous motivation to engage in work activities that contribute to the employees’ professional development (Ryan and Deci, 2017) and that of others (Koopmans et al., 2011), which can also contribute to the goals of the organization.

286  Research handbook on human resource management and disruptive technologies

THEORETICAL FRAMEWORK AND HYPOTHESES Contextual Work Performance Contextual work performance can be defined as those activities that enhance the effectiveness of the organization through its positive effects on the psychological, social and organizational work context (Motowidlo et al., 1997; Koopmans et al., 2011). Contextual work performance differs from task performance, which refers to the skill or ability to perform the core or central tasks of the job (Koopmans et al., 2011). It refers to those employee behaviors and actions that are developed outside of the employees’ main tasks and that support the organization, such as investing in new knowledge and skill development to perform additional tasks, taking on challenging extra tasks, taking the initiative, and leading and developing others (Koopmans et al., 2011). Alternative labels for this concept that also relate to doing extra tasks, displaying initiative and socializing with others are nonjob-specific task proficiency, extra-role performance, organizational citizenship behavior and interpersonal relations (see Koopmans et al., 2011 for an overview). Contextual work performance indirectly contributes to the performance of the organization as it facilitates task execution. For example, employees’ contextual work performance, including proactive and creative performance, can influence others so that they are more likely to exhibit behavior that contributes to the effectiveness of the organization (Koopmans et al., 2011). The Relationship between Collegial Isolation and Contextual Work Performance The concept of professional isolation reflects employees’ desire to be part of a network of colleagues within an organization who collectively provide help and support with specific work-related needs and represents employees’ perceptions about the availability of colleagues and supervisors and their support (Marshall et al., 2007). Within the scientific literature, professional isolation is often described as a perception of a lack of required resources, distance from colleagues and the workplace or issues related to work. Bedward and Daniels (2005), for example, positioned professional isolation in a social context, referring to a feeling of not being supported, having no (career or job) opportunities and not being recognized or praised for achievements. Others described professional isolation as a lack of a sense of support from colleagues (Bushy, 2004), a lack of communication (Vimarlund et al., 2008) or a lack of experiencing mentorship (Stewart and Carpenter, 2009). The above descriptions focus on both the organization and colleagues. Professional isolation, therefore, is presented as a two-fold construct (Marshall et al., 2007; Orhan et al., 2016). The two-dimensional construct of Marshall and colleagues (2007) consists, on the one hand, of the perception of isolation in relation to the relatedness with the organization and belongingness to the supporting network within this organization (organizational isolation) and, on the other hand,

Keep in touch in remote workplaces  287

of employees’ perceptions of being isolated in relation to the physical meeting of colleagues (collegial isolation) (Marshall et al., 2007). Both dimensions reflect the desire of employees to be ‘to be part of the network of colleagues who provide help and support in specific work-related needs. It represents employees’ perceptions of availability of co-workers, peers, and supervisors for work-based social support’ (Marshall et al., 2007, p. 199). Since the number of physical meetings with colleagues has dropped drastically since the COVID-19 pandemic, the focus in this study is on collegial isolation. Research by Pinsonneault and Boisvert (2001) showed that the degree of face-to-face interactions has a strong influence on employees’ perceptions of the social interactions and support received from peers in the workplace. Collegial isolation refers to the lack of proximity in relation to the peers that individuals can go to, discuss their daily work-related problems with, and that can be considered their friends at work (Marshall et al., 2007). Collegial isolation, therefore, can be considered an important antecedent of contextual work performance. In view of this, we propose the following: Hypothesis 1: Collegial isolation is directly and negatively related to contextual work performance. The Mediating Role of Relatedness in the Relationship between Peer Isolation and Contextual Work Performance The theory of self-determination (SDT) assumes that fundamental psychological satisfaction of the three basic psychological needs (i.e. autonomy, competence and relatedness) represents the underlying motivating mechanism that stimulates and directs people’s autonomous behavior (Deci and Ryan, 2000), also regarding work. The satisfaction of these basic psychological needs is, therefore, seen as an important predictor of employees’ optimal functioning (Van den Broeck et al., 2010), including their motivation to display contextual work performance. Particularly the need for relatedness, however, which refers to the individual’s natural desire to being connected and being associated with other people (Deci and Ryan, 2000), might be compromised or under pressure due to the mandatory working from home during the COVID-19 pandemic. During the COVID-19 pandemic, due to intensified tele(home)working, it was harder for employees to build up a (close) bond with each other and to have the feeling of being involved and part of a group. There was little opportunity for being among others and interacting with each other so as to build positive relationships and to feel that one is cared for and can care about others. In particular, the feeling of involvement with a group facilitates internalization of values and behaviors that are endorsed in that environment (Deci and Ryan, 2008). Also, contextual behavior can be expected to be prompted, modeled and valued by significant others to whom employees feel (or want to feel) attached or related (Deci and Ryan, 2000). Hence, employees who experience relatedness are more likely to engage in contextual work performance, as they may want to positively influence

288  Research handbook on human resource management and disruptive technologies

others to adopt the same behavior and thus internalize it and to be more effective in daily work (Motowidlo et al., 1997). Therefore, we propose the following: Hypothesis 2: Relatedness is directly and positively related to contextual work performance. Employees experience feelings of relatedness when they are part of a team and can share personal feelings and thoughts with colleagues. The assumption that individuals have a natural tendency to seek out this feeling and integrate themselves into social circles and benefit from caring about them is emphasized in the attachment theory (Van den Broeck et al., 2010). In addition, it is consistent with concepts in organizational psychology, such as social support (Viswesvaran et al., 1999) and loneliness at work (Wright et al., 2006; Wright and Silard, 2020). Collegial isolation concerns the lack of integration into social circles and the lack of colleagues who can be considered as friends and who care about each other (Marshall et al., 2007). Collegial isolation reflects that an employee feels that he or she has built a less close bond with their colleagues and is not involved in or part of a group. This leads to the following hypothesis: Hypothesis 3: Collegial isolation is directly and negatively related to belongingness. Summarizing, it can be stated that collegial isolation is expected to affect contextual performance both directly and indirectly through relatedness, which functions as a mediator. A lower perception of relatedness that stems from more collegial isolation can be expected to be associated with lower contextual work performance. This leads to the following: Hypothesis 4: Relatedness mediates the negative relationship between collegial isolation and contextual work performance.

METHODOLOGY Sample Data were collected by means of a self-report questionnaire that was distributed via online channels, including personal network accounts, such as LinkedIn and Facebook, communication tools, such as e-mail and WhatsApp, and through personal individual contact with respondents. A total of 216 respondents completed the questionnaire. Based on listwise deletion, the answers of nineteen respondents were removed due to the lack of answers. After removing the respondents with missing data, 177 respondents remained. From these 177 respondents, 146 were salaried employed, which were used for further analysis. To provide insight into the collective characteristics of the respondents used in the study, the descriptive statistics of the sample are presented in Table 20.1.

Keep in touch in remote workplaces  289

Table 20.1

Overview sample

 

Contextual Work Performance N

%

Mean

SD

Gender

 

Male

85

58.2

2.90

0.78

Female

60

41.1

3.04

0.90

Unknown/missing

1

0.7

 

 

Age categories

 

0.07) (cf. Hair et al., 2014). To have enough construct validity the average variance extracted (AVE), which needs to exceed the value of 0.50 (Fornell and Larcker, 1981), was calculated. One item was deleted from the collegial isolation scale to have adequate reliability and convergent validity. Also, one item was deleted from the contextual work performance scale for the same purpose. No items were deleted from the relatedness scale. We checked for discriminant validity by examining and comparing the AVEs of each respective construct with the interconstruct correlations in the model to investigate whether each latent variable shared greater variance with its own measurement items than with the other constructs (Fornell and Larcker, 1981; Chin, 1998). Our analysis demonstrated that the model has sufficient discriminant validity. An overview of the reliability and validity checks is presented in Table 20.2. We subsequently examined indicator reliability. All factor loadings were above 0.60 and therefore acceptable (Hair et al., 2014). Finally, we checked for discriminant validity, comparing the AVEs of the constructs with the interconstruct correlations determining whether each latent variable shared greater variance with its own

Keep in touch in remote workplaces  291

Table 20.2  

Construct descriptive statistics N

Theoretical

Actual Range

Mean

SD

Range

Cronbach’s

AVE

Alpha

Collegial Isolation 146

1.00–5.00

1.60–5.00

3.68

0.81

0.86

0.66

Contextual Work

146

1.00–5.00

1.00–4.86

3.01

0.84

0.84

0.50

146

1.00–5.00

1.33–4.67

3.30

0.73

0.84

0.56

Performance Relatedness

Table 20.3

Correlations second wave and the square root of the Average Variance Extracted (in bold)

 

Collegial Isolation

Contextual Work

Relatedness

Performance Collegial Isolation

0.81

Contextual Work Performance

-0.34**

  0.71

Relatedness

-0.73**

0.28**

    0.75

Note: Significance correlations: ** = p < 0.01.

measurement variables or with other constructs (Fornell and Larcker, 1981; Chin, 1998). We compared the square root of the AVE for each construct with the correlations with all other constructs in the model (Table 20.3). A correlation between constructs exceeding the square roots of their AVEs indicates that they may not be sufficiently discriminable. For each construct, we found that the absolute correlations did not exceed the square roots of the AVEs. Hence, we concluded that all constructs showed sufficient reliability and validity. Model Estimates Bootstrap t-statistics were used for testing the significance of the path-coefficients (Anderson and Gerbing, 1988). An estimation of model fit was made with a standardized root mean square residual (SRMR), showing an adequate model fit value of 0.08 for the model, which meets the criterion set by Hu and Bentler (1998). Hypothesis Testing We tested the direct effects as listed in the first three hypotheses. Finally, we tested the mediation effect of relatedness in the model (Hypothesis 4) by calculating the indirect effects of collegial isolation via relatedness on contextual work performance. The results are summarized in Table 20.4. As depicted in Table 20.4, Hypothesis 1 is supported; a negative relationship was found between collegial isolation and contextual work performance (γ = –0.29, p < 0.01, R2 = 0.11). Hypothesis 2 was not supported; no significant relationship was found between relatedness and contextual work performance (γ = 0.06, p < 0.51).

292  Research handbook on human resource management and disruptive technologies

Table 20.4

Direct and indirect effects

 

Coefficient (γ)

SD

T Statistics

P Values

Hypotheses

Collegial Isolation

–0.29

0.11

2.56

0.01

H1

–0.73

0.04

18.26

0.00

H3

0.06

0.13

0.51

0.61

H2, H4

→ Contextual Work Performance Collegial Isolation → Relatedness Relatedness → Contextual Work Performance

Hypothesis 3 was supported; a negative significant relationship was found between collegial isolation and relatedness (γ = –0.73, p < 0.00, R2 = 0.53). Hypothesis 4, however, was not supported; no significant relationship was found in the indirect effect of collegial isolation via relatedness on contextual work performance (γ = –0.046, p < 0.62).

CONCLUSION AND RECOMMENDATIONS FOR ACTION The aim of this research was to contribute to the scientific and societal debates about professional isolation in telework contexts, which have been rekindled since the start of the COVID-19 pandemic, which forced organizations, employees and other stakeholders to adopt intensive and enduring remote work practices (telework) enabled by disruptive information and communication technologies, such as Skype, Microsoft Teams and Zoom. Based on the telework literature, the performance literature and self-determination theory (SDT), we tested a set of hypotheses via PLS-SEM analysis, employing data collected during the COVID-19 pandemic from 146 salaried employees. The main results and their implications are discussed below. Discussion of the Main Findings First, in line with expectations, this study revealed a direct and negative relationship between collegial isolation and contextual work performance, referring to work-related tasks beyond an employee’s task description and that can contribute to the individual’s, colleagues’ and organizational performance. This result is in line with the results of Marshall et al. (2007) who suggested that experiencing professional isolation can be associated with lower work performance. The negative effect of collegial isolation on contextual work performance can be considered as an unpleasant experience (Shaver and Brennan, 1991). On the one hand, therefore, collegial isolation can be characterized as a work stressor that has negative effects on performance (Weiss, 1975; Toscano and Zappalà, 2020), as it decreases employees’ confidence in their abilities and knowledge (Marshall et al., 2007; Golden et al., 2008). On the other hand, collegial isolation can also be taken to reflect a lack of

Keep in touch in remote workplaces  293

social support, an important job demand in the job demands-resources model (Bakker and Demerouiti, 2007; Sardeshmukh et al., 2012), reducing employees’ work engagement (i.e. vigor, dedication and absorption). The lack of this job resource, therefore, can hinder employees’ contextual performance (Schaufeli, 2018). In both perspectives, collegial isolation reflects a lack of interaction with others, which is an important factor in contextual work performance, such as taking on new tasks and responsibilities, keeping up with knowledge and skills and taking on new challenges (Koopmans et al., 2011). The negative relationship between collegial isolation and contextual work performance can be taken in two ways. On the one hand, the lack of interaction with others reflects employees missing the support of, and the motivation to mirror the contextual behavior of, others. On the other hand, the lack of interaction with others reflects employees not being motivated to inspire others by showcasing contextual work behavior (Marshall et al., 2007; Koopmans et al., 2011, 2014). Second, also in line with expectations, we found a negative relationship between social isolation and relatedness. Based on SDT, we assumed that collegial isolation would limit socialization and, therefore, an employee’s feeling of being part of a larger group, reflected in the notion of relatedness (Deci and Ryan, 2000), one of the three basic psychological needs (i.e. autonomy, competence and relatedness) that can be satisfied in the work domain and motivate autonomous workplace behavior. Third and fourth, based on SDT, we expected that employees would be autonomously motivated to perform contextual work behavior, as this is valued by significant others (e.g. colleagues, bosses, clients) in a group they feel related to or feel they belong to (Deci and Ryan, 2000). In contrast to this expectation, however, we did not find a significant direct relationship between relatedness and contextual performance. Hence, also in contrast to our expectations, no support was found for relatedness being a mediator in the relationship between collegial isolation and contextual performance Although prior SDT-research suggested that the satisfaction of relatedness would be an important predictor for the optimal functioning of individuals and can contribute to the internalization of values and behaviors (Deci and Ryan, 2008; Van den Broeck et al., 2010), particularly in remote work contexts the feeling of connectedness as such might not be sufficient to engage in the valued contextual work performance. However, since collegial isolation was found to be directly related to contextual performance, it may be concluded that the perceived availability of colleagues and the concrete (physical) interactions with them and the support they provide are important elements that can motivate contextual work behavior. In this regard, our study extends the SDT-theory by demonstrating that (in remote work contexts) the satisfaction of employees’ need for relatedness is not sufficient for stimulating employees’ contextual work behaviors. Rather the perceived availability of and interaction with (perhaps only a few) colleagues one has an affective bond with may play a role in employees’ contextual work performance. These can provide emotional support and allow the employee to mirror their thoughts and behaviors (Marshall et al., 2007; Golden et al., 2008). It makes it possible to have informal chats and spontaneous discussions (Cooper and Kurland, 2002), which can inspire and support the employee in disruptive work contexts, enhancing one’s confidence

294  Research handbook on human resource management and disruptive technologies

in one’s abilities, or reversely. Alternatively, having close interactions can also motivate employees to inspire and coach others (Koopmans et al., 2011). Limitations and Avenues for Future Research Despite its contributions, the current study has several limitations. First, this study focused on collegial isolation only, representing employees’ perceptions of isolation due to a lack of social interactions with colleagues. However, the perception of isolation from the organization or from other stakeholders, implying a lack of work-based support from one’s supervisor and the organization, or from other stakeholders inside or outside the organization, was not included. In increasingly hybrid work environments, however, not only collegial isolation, but also isolation from other parties is likely. Yang et al. (2021), for example, showed that tele(home)working during the COVID-19 pandemic affected employees’ collaboration with others beyond team boundaries, perhaps leading to a wider feeling of professional isolation. Future research could include employees’ perceptions of not being a member of the team or the departmental network and of not being acknowledged by the organization (Marshall et al., 2007), or other relevant parties outside the organization. Also this type of isolation can hinder individual, team, organizational and societal performance. Second, the study focused on contextual performance, which reflects important employee activities and behaviors that help to support and develop employees themselves, others and the organization, which is particularly important in disruptive work environments. In disruptive work environments, however, task performance (being indicated in the job description and part of the employees’ work routines) and counterproductive performance may be affected by relatedness, as employees do not need to interact to engage in these. Moreover, another important dimension of individual performance in disruptive work environments, which was not explicitly included in the study, but can be considered part of contextual work performance, is adaptive performance. This type of performance can be defined as ‘the extent to which an individual adapts to changes in a work system or work roles. It includes, for example, solving problems creatively, dealing with uncertain or unpredictable work situations, learning new tasks, technologies, and procedures, and adapting to other individuals, cultures, or physical surroundings’ (Koopmans et al., 2011, p. 862). Future research in disruptive contexts could focus on the relationship between professional isolation and other forms of individual performance, and in particular adaptive performance. Third, we focused on the mediating role of relatedness as an important basic psychological need that must be fulfilled in disruptive contexts to stimulate autonomous employee behavior that can both support one’s own professional development and that of others and the organization. We did not focus on other basic psychological needs, such as the need for autonomy and competence (Deci and Ryan, 2000). These two needs, however, can also play a role in disruptive situations, as they enhance work engagement and hence promote proactive work behavior (Schaufeli, 2018).

Keep in touch in remote workplaces  295

Fourth, since we used a cross-sectional design, we are unable to comment on the dynamic interplay between professional (collegial) isolation, satisfaction of the basic psychological need for relatedness, and individual (contextual) work performance. Likely, relatedness can reduce professional isolation in the longer run, contextual work performance can enhance relatedness and reduce professional isolation, et cetera. We encourage future researchers, therefore, to conduct longitudinal study-designs to disentangle these interrelationships. Fifth, the present study only focused on employees’ experiences during the COVID-19 pandemic in which tele(home)working was enforced, as an example of a disruptive condition at work. Experiences and preferences, for example, regarding tele(home)working, may have shifted after the pandemic. Hence, future studies could focus on other disruptive and perhaps remote work environments, caused by other factors such as technological developments. Sixth, no explicit attention was paid to the role of the disruptive technologies in the relationships under study. However, technology, particularly the quality thereof in communication and interaction processes in remote workplaces, can play an important role in reducing professional isolation by enhancing the employees’ presence awareness while communicating via ICT (Malhotra and Majchrzak, 2014). Management Implications Our findings have some practical implications for HR staff and departments, managers and employees in a disruptive (distributed) work environment or hybrid work context. First, to enhance employees’, colleagues’ and organizational development and performance, stimulating interactions is more important than the general feeling of relatedness one experiences with a wider group of colleagues, bosses, subordinates, clients, suppliers, the organization or the community at large. Based on this, companies and HR managers can be advised to develop policies to prevent professional (collegial) isolation, as this is shown to be associated with lower contextual performance. Second, even in contexts that have been disrupted by (enforced) remote working, measures need to be taken to maintain individual (contextual) work performance, to improve or to prevent a decline in professional and organizational development. Colleagues can also be more aware of the potential consequences of professional isolation and the effect and role they can play in this. Based on this, they can take action to counteract any negative effects. Finally, the same applies to the individual employee. Moreover, even though relatedness, as a subdimension of self-determination, was not found to affect contextual behavior, it can influence other types of individual performance, Therefore, individuals, managers and the organization should be active in maintaining mutual relationships.

296  Research handbook on human resource management and disruptive technologies

REFERENCES Achor, S., Rosen Kellerman, G., Reece, A., and Robichaux, A. (2018). America’s loneliest workers, according to research. Harvard Business Review, 11 April. https://​hbr​.org/​2018/​ 03/​americas​-loneliest​-workers​-according​-to​-research. Allen, T. D., Golden, T. D., and Shockley, K. M. (2015). How effective is telecommuting? Assessing the status of our scientific findings. Psychological Science in the Public Interest, 16(2), 40–68. https://​doi​.org/​10​.1177/​1529100615593273. Anderson, D. and Kelliher, C. (2020). Enforced remote working and the work-life interface during lockdown. Gender in Management: An International Journal, 35(7/8), 677–83. https://​doi​.org/​10​.1108/​gm​-07​-2020​-0224. Anderson, J. C. and Gerbing, D. W. (1988). Structural equation modeling in practice: a review and recommended two-step approach. Psychological Bulletin, 103(3), 411–23. https://​doi​ .org/​10​.1037/​0033​-2909​.103​.3​.411. Bakker, A. B. and Demerouti, E. (2007). The job demands-resources model: state of the art. Journal of Managerial Psychology, 22(3) 309–28. Bedward, J. and Daniels, H. R. J. (2005). Collaborative solutions – clinical supervision and teacher support teams: reducing professional isolation through effective peer support. Learning in Health and Social Care, 4(2), 53–66. https://​doi​.org/​10​.1111/​j​.1473​-6861​.2005​ .00090​.x. Borman, W. C. and Motowidlo, S. J. (1997). Task performance and contextual performance. The meaning for personnel selection research. Human Performance, 10(2), 99–109. https://​ doi​.org/​10​.1207/​s15327043hup1002​_3. Bushy, A. (2004). Creating nursing research opportunities in rural healthcare facilities. Journal of Nursing Care Quality, 19(2), 162–8. https://​doi​.org/​10​.1097/​00001786​-200404000​ -00014. Chin W. W. (1998). The partial least squares approach to structural equation modeling. In G. A. Marcoulides (ed.), Modern Methods for Business Research (pp. 295–336). Lawrence Erlbaum Associates. Cooper, C. D. and Kurland, N. B. (2002). Telecommuting, professional isolation, and employee development in public and private organizations. Journal of Organizational Behavior, 23(4), 511–32. https://​doi​.org/​10​.1002/​job​.145. Deci, E. L. and Ryan, R. M. (2000). The ‘what’ and ‘why’ of goal pursuits: human needs and the self-determination of behavior. Psychological Inquiry, 11(4), 227–68. https://​doi​.org/​10​ .1207/​s15327965pli1104​_01. Deci, E. L. and Ryan, R. M. (2008). Facilitating optimal motivation and psychological well-being across life’s domains. Canadian Psychology/Psychologie Canadienne, 49(1), 14–23. https://​doi​.org/​10​.1037/​0708​-5591​.49​.1​.14. Fornell, C. and Larcker, D. F. (1981). Structural equation models with unobservable variables and measurement error: algebra and statistics. Journal of Marketing Research, 18(3), 382–8. https://​doi​.org/​10​.2307/​3150980. Galanti, T., Guidetti, G., Mazzei, E., Zappalà, S., and Toscano, F. (2021). Work from home during the COVID-19 outbreak: the impact on employees’ remote work productivity, engagement, and stress. Journal of Occupational and Environmental Medicine, 63(7), e426–e432. https://​doi​.org/​10​.1097/​JOM​.0000000000002236. Golden, T. D. and Gajendran, R. S. (2018). Unpacking the role of a telecommuter’s job in their performance: examining job complexity, problem solving, interdependence, and social support. Journal of Business and Psychology, 34(1), 55–69. https://​doi​.org/​10​.1007/​s10869​ -018​-9530​-4. Golden, T. D., Veiga, J. F., and Dino, R. N. (2008). The impact of professional isolation on teleworker job performance and turnover intentions: does time spent teleworking, interact-

Keep in touch in remote workplaces  297

ing face-to-face, or having access to communication-enhancing technology matter? Journal of Applied Psychology, 93(6), 1412–21. https://​doi​.org/​10​.1037/​a0012722. Hair J. F., Sarstedt, M., Hopkins, L., and Kuppelwieser, V. G. (2014). Partial least squares structural equation modeling (PLS-SEM) an emerging tool in business research. European Business Review, 26(2), 106–21. https://​doi​.org/​10​.1108/​EBR​-10​-2013​-0128. Hamersma, M., De Haas, M., and Faber, R. (2020). Working from home and the corona crisis. Knowledge Institute for Mobility Policy, August. https://​www​.kimnet​.nl/​publicaties/​ rapporten/​2020/​08/​31/​thuiswerken​-en​-de​-coronacrisis. Hu, L. T. and Bentler, P. M. (1998). Fit indices in covariance structure modeling: sensitivity to under-parameterized model misspecification. Psychological Methods, 3(4), 424–53. https://​ doi​.org/​10​.1037/​1082​-989X​.3​.4​.424. Koopmans, L., Bernaards, C. M., Hildebrandt, V. H., Schaufeli, W. B., De Vet , H. C., and Van der Beek, A. J. (2011). Conceptual frameworks of individual work performance. Journal of Occupational & Environmental Medicine, 53(8), 856–66. https://​doi​.org/​10​.1097/​jom​ .0b013e318226a763. Koopmans, L., Bernaards, C., Hildebrandt, V., De Vet, R., and Van der Beek, A. (2014). The individual work performance questionnaire (IWPV): internal consistency, construct validity and standardization. Journal of Health Sciences, 92(6), 231–9. https://​doi​.org/​10​.1007/​ s12508​-014​-0090​-3. Kurland, N. B. and Bailey, D. E. (1999). The advantages and challenges of working here, there anywhere, and anytime. Organizational Dynamics, 28(2), 53–68. doi:10.1016/ s0090-2616(00)80016-9. Malhotra, A. and Majchrzak, A. (2014). Enhancing performance of geographically distributed teams through targeted use of information and communication technologies. Human Relations, 67(4), 389–411. Marshall, G. W., Michaels, C. E., and Mulki, J. P. (2007). Workplace isolation: exploring the construct and its measurement. Psychology and Marketing, 24(3), 195–223. https://​doi​.org/​ 10​.1002/​mar​.20158. Motowidlo, S. J., Borman, W. C., and Schmit, M. J. (1997). A theory of individual differences in task and contextual performance. Human Performance, 10(2), 71–83. https://​doi​.org/​10​ .1207/​s15327043hup1002​_1. Mulki, J. P., Locander, W. B., Marshall, G. W., Harris, E. G., and Hensel, J. (2008). Workplace isolation, salesperson commitment, and job performance. Journal of Personal Selling & Sales Management, 28(1), 67–78. https://​doi​.org/​10​.2753/​pss0885​-3134280105. Orhan, M. A., Rijsman, J. B., and Van Dijk, G. M. (2016). Invisible, therefore isolated: comparative effects of team virtuality with task virtuality on workplace isolation and work outcomes. Revista de Psicología del Trabajo y de las Organizaciones, 32(2), 109–22. https://​ doi​.org/​10​.1016/​j​.rpto​.2016​.02​.002. Orr, J. M., Sackett, P. R., and Mercer, M. (1989). The role of prescribed and nonprescribed behaviors in estimating the dollar value of performance. Journal of Applied Psychology, 74(1), 34–40. https://​doi​.org/​10​.1037/​0021​-9010​.74​.1​.34. Pinsonneault, A. and Boisvert, M. (2001). The impacts of telecommuting on organizations and individuals: a review of the literature. In N. J. Johnson (ed.), Telecommuting and Virtual Offices: Issues and Opportunities (pp. 163–85). Idea Group Publishing. Ringle, C. M., Wende, S., and Becker, J. M. (2015). SmartPLS3. SmartPLS Gmbh. http://​www​ .smartpls​.com. Ryan, R. M. and Deci, E. L. (2017). Self-determination theory: Basic psychological needs in motivation, development, and wellness. The Guilford Press. Sardeshmukh, S. R., Sharma, D., and Golden, T. D. (2012). Impact of telework on exhaustion and job engagement: a job demands and job resources model. New Technology, Work and Employment, 27(3), 193–207.

298  Research handbook on human resource management and disruptive technologies

Schaufeli, W. B. (2018). Work engagement in Europe: relations with national economy, governance and culture. Organizational Dynamics, 47(2), 99–106. https://​doi​.org/​10​.1016/​ j​.orgdyn​.2018​.01​.003. Shaver, P. R. and Brennan, K. A. (1991). Measures of depression and loneliness. Measures of Personality and Social Psychological Attitudes, 6(11), 195–289. https://​doi​.org/​10​.1016/​ b978​-0​-12​-590241​-0​.50010​-1. Stewart, S., and Carpenter, C. (2009). Electronic mentoring: an innovative approach to providing clinical support. International Journal of Therapy and Rehabilitation, 16(4), 199–206. https://​doi​.org/​10​.12968/​ijtr​.2009​.16​.4​.41191. Toscano, F. and Zappalà, S. (2020). Social isolation and stress as predictors of productivity perception and remote work satisfaction during the COVID-19 pandemic: the role of concern about the virus in a moderated double mediation. Sustainability, 12, 9804. https://​ doi:​10​.3390/​su12239804. Van den Broeck, A., Vansteenkiste, M., De Witte, H., and Lens, W. (2008). Explaining the relationships between job characteristics, burnout, and engagement: the role of basic psychological need satisfaction. Work & Stress, 22(3), 277–94. https://​doi​.org/​10​.1080/​ 02678370802393672. Van den Broeck, A., Vansteenkiste, M., Witte, H., Soenens, B., and Lens, W. (2010). Capturing autonomy, competence, and relatedness at work: construction and initial validation of the work-related basic need satisfaction scale. Journal of Occupational and Organizational Psychology, 83(4), 981–1002. https://​doi​.org/​10​.1348/​096317909x481382. Van Veldhoven, M. and Van Gelder, M. (2020). The advantages and disadvantages of mandatory working from home during the lockdown. Tijdschrift voor HRM, 23(3), 66–90. https://​ doi​.org/​10​.5117/​thrm2020​.3​.veld. Vimarlund, V., Olve, N. G., Scandurra, I., and Koch, S. (2008). Organizational effects of information and communication technology (ICT) in elderly homecare: a case study. Health Informatics Journal, 14(3), 195–210. https://​doi​.org/​10​.1177/​1081180x08092830. Viswesvaran, C., Sanchez, J. I., and Fisher, J. (1999). The role of social support in the process of work stress: a meta-analysis. Journal of Vocational Behavior, 54(2), 314–34. https://​doi​ .org/​10​.1006/​jvbe​.1998​.1661. Weiss, R. S. (1975). Loneliness: The Experience of Emotional and Social Isolation. MIT Press. Wiesenfeld, B. M., Raghuram, S., and Garud, R. (2001). Organizational identification among virtual workers: the role of need for affiliation and perceived work-based social support. Journal of Management, 27(2), 213–29. Wright, S. L., Burt, C. D., and Strongman, K. T. (2006). Loneliness in the workplace: construct definition and scale development. New Zealand Journal of Psychology, 35(2), 59–68. https://​www​.researchgate​.net/​publication/​29487439​_Loneliness​_in​_the​_Workplace​ _Construct​_definition​_and​_scale​_development. Wright, S. and Silard, A. (2020). Unravelling the antecedents of loneliness in the workplace. Human Relations, 74(7), 1060–81. https://​doi​.org/​10​.1177/​0018726720906013. Yang, L., Holtz, D., Jaffe, S., Suri, S., Sinha, S., Weston, J., and Teevan, J. (2021). The effects of remote work on collaboration among information workers. Nature Human Behaviour, 6, 1–12. https://​doi​.org/​10​.1038/​s41562​-021​-01196​-4. Yerkes, M. A., Remery, C., André, S., Salin, M., Hakovirta, M., and Van Gerven, M. (2020). Unequal but balanced: highly educated mothers’ perceptions of work–life balance during the COVID-19 lockdown in Finland and the Netherlands. Journal of European Social Policy, 09589287221080411.

Index

machine learning 193, 194, 196, 201 proxies in 41 surveillance 26, 30 vigilance of effects of 66 AM see algorithmic management (AM) Amauger-Lattes, M. C. 77 AMS see Austrian ArbeitsMarktService (AMS) Anderson, J. 11 Aneesh, A. 60 Argyris, C. 78, 84–5 Arrowsmith, James 39 artificial intelligence (AI) 39–40, 193, 224, 260–61 adoption 13–14, 19 augmentation 226, 228–31, 234 automation 226, 228–31, 234 bias mitigation techniques, challenges in 45–6 China’s regulations 69 coexistence of people and 226–7 job role changes 231–5 paradox theory 228–9 patterns of 230–31 computational bias 41–2 defined 225 in e-HRM research 136–7 applicant’s personality 146 automation of recruitment 145, 147 bibliometric analysis (see bibliometric analysis, in AI and HRM) recruiting and selection 143–7 training and performance evaluation 146 EU regulation 68, 76 and future of jobs 9–10, 15, 19 automation 10, 13, 16 disruptions 10–14 polarized debates 10–14, 20 and future of work 18 governance of 75 in hiring process 42, 52–3, 55, 57, 68 assessment 200–201 outputs 199–200, 202 for HRM 39–40

Abrahamson, E. 9 accountability 64, 66, 154, 159 Acee, H. J. 156 adaptive structuration theory (AST) 107, 111 affordances developmental vs. low-cost labour sourcing 244–6 marketplace vs. community 246–7 new role vs. traditional role 247–9 agency conflicts 215, 220–21 end-user 214–15 technological 212 theory 207–10 digi-generations and 210–12 technology and 209–10 AI see artificial intelligence (AI) AI hiring tool 193–6, 198–202 data collection and analysis 196–7 in unconscious bias training 199 Alavi, M. 108 algocracy 60 algorithmic bias, human bias vs. 66–7 algorithmic management (AM) 23–5, 74–7, 87, 96–8, 101 autonomy needs 26–7, 29 competence needs 26–9 defined 75 fault line 75, 86 functions 23–4 motivation-enhancing 28–33 regulation 76 relatedness needs 26, 27, 30 trade unions 85–6 algorithms/algorithmic 39, 41, 42–6, 52–6 aversion 52 decision-making 29, 62, 64, 173 digital 60 digitalized talent management 173 discriminatory elements 53 for facial recognition 41 feedback 27, 33 hiring 41–2, 44, 54, 195, 197–202 HRM 24, 145 and humans 60–62 leadership 24 299

300  Research handbook on human resource management and disruptive technologies

bias in developing AI 40–43 decisions 39 impact on jobs 18 research questions 15–17 innovations 18, 20 mitigating biases 43–5 mundane and incremental nature of 230 opacity of 261 as opportunity for augmentation 17 in organizational settings 225–6 polarized debates 10–14, 20 power of 16 and robotics 11, 15 solving the labour market shortages 18, 20 standardization 69 statistical bias 41–2 systematic bias 40–41 technical constraints 66 technological disruptions 12 US regulation 68–9 work management 39 Artificial Intelligence Act 57, 68–71, 76 AST see adaptive structuration theory (AST) Atkinson, R. D. 15 attachment theory 288 augmentation 16, 17, 226, 227 artificial intelligence 226, 228–31, 234 paradox theory of 227–8 Austrian ArbeitsMarktService (AMS) 52–3 authoritarian leadership 111 automatic decision process 62 automation 13, 16, 18, 62, 136, 145–6, 167, 210, 226, 227 artificial intelligence 226, 228–31, 234 digitalization in 169–70 hiring process 62, 145–7 jobs 10 risk of 10 autonomy needs 26–7, 29 of worker 258, 260–61 in workplace 64 average variance extracted (AVE) 290–91 Avolio, B. J. 107, 109, 110, 113 Bader, P. 108 Bainbridge, S. 16 Baiocco, S. 74 Ball, M. 85 Bateson, G. 84–5 Baumruk, R. 153

Bauwens, R. 111 Beane, M. 263 Bedward, J. 286 Belitski, M. 114 Bentler, P. M. 291 Bergmann, J. 82 Berkers, H. A. 260–65 biases 199 in algorithms 173 computational 40–42 in developing AI for HRM 40–54 human 42–3 mitigating AI 43–6 recruiting systems 62 selection/exclusion 42–3 statistical 40–42 systemic 40–41 bibliometric analysis, in AI and HRM 137–8, 146–8 cluster analysis 143, 147, 148 co-words analysis 138, 143 data collection and analysis 138–9 keywords analysis 142–3, 147 most productive authors 139 publication by year 139 publishing activity by country 139–41 relevant topics 141–2 thematic map analysis 143–7 big data 125 Boettcher, K. 112 Boisvert, M. 287 Bondarouk, T. 24, 75 bots 64–5 Boudreau, M. C. 209, 210 bounded automation 257 Bowen, D. E. 94, 95 Brandes, P. 13 Brandi, U. 81 Breidensjö, M. 85 Brown, J. S. 79 Brynjolfsson, E. 15, 16 Buch, A. 81 Bujold, A. 23 Burgoyne, J. 85 career 247 Carter, B. 84 Carte, T. A. 108 changing work 270, 274 Charbonneau, É. 30 Cheng, X. 207 Christensen, C. M. 156

Index  301

Clarke, A. E. 82 cluster analysis 143, 147, 148 Cobo, M. J. 143 cobot technology 254–6, 264 advantage 263 cognitive demands, reduction of 262 human-robot interaction 255–6 impact on skills 264 Industry 4.0 256, 264 negative effect on task variety 261 technological functionality 260 co-creation-oriented HRM system 122–5 codes of conduct 56 coexistence concept 226 of people and AI 226–7 job role changes 231–5 paradox theory 228–9 Cohen, Y. 254 collaboration human-cobot 255–6, 265 on cognitive demands 262 impact on job characteristics 258–64 incorporation of artificial intelligence 260–61 mental workload 263 research 255–6 threat to the worker’s autonomy 258, 261–2 collaborative technologies 256–7 collective bargaining agreements 56, 76–7 collective participation 78–9 collegial isolation 284, 285, 287, 293, 295 and contextual work performance, relationships 286–7, 292–3 hypothesis testing 291–2 limitations and future research 294 measure of 289–90 model characteristics 290 negative effect of 292 Collings, D. G. 178 communication 154, 159 communities of practice (CoP) 79–80, 82, 85 community manager 242, 246, 248, 249 compensation and benefits 124 competence needs 26–9 competition 31 computational bias 40–42 conceptual structure map 143–7 configuration services, internet-enabled 126–7

contextual work performance 284–8, 293, 295 collegial isolation and, relationships 286–7 hypothesis testing 291–2 limitations and future research 294 measure of 289 model characteristics 290 Cooper, L. 85 CoP see communities of practice (CoP) Cordery, J. 108 corporate governance 58–9, 64 Cortellazzo, L. 111, 113 counterproductive work performance 284, 289, 294 Cousins, K. C. 209 COVID-19 pandemic and digitalization 168–9 drop of training and development 275 great resignation in 274 inequalities in working life 275 remote work in 272, 274, 284, 285, 287, 292 and risk of collegial isolation 284 tele-home working 285, 287, 294, 295 Cowan, L. D. 109 co-words analysis 138, 143 creative destruction 156 Creswell, J. W. 213 customer abilities 127, 129 attributes 120, 122, 128, 129 development 124 engagement 124–5 motivation 126, 129 opportunities 128, 129 ratings 31 role readiness 122 customer value 120, 122, 123 customer abilities on 127 HRM systems and 121, 125–9 Cyert, R. M. 77 Danaher, J. 60 Daniels, H. R. J. 286 datasets 40, 44 computational bias in 41–2 human bias in 43 statistical bias 40–42 systemic biases in 40–41 Davenport, T. H. 16, 17, 168 DeChurch, L. A. 108

302  Research handbook on human resource management and disruptive technologies

decision-making 43, 44, 63, 64, 145, 167, 168, 195 algorithmic 29, 62, 64, 173 digitalization 169 in hiring 66, 196, 201–2 HRM 170–71 judgments of value 172, 174 subjective perceptions 174 talent management 170–74 in technology 169–70 Desouza, K. C. 208 De Stefano, V. 76 Dewey, J. 80–81 digi-generations 210–12, 220–21 digital tools use of 215–20 online interviews 212–20 digital assets 59 bartering 64 ethics 60 inequality 211 interface 161 labour platforms 55, 238, 239 leadership 105, 106, 110–13 media 232 platforms 11 platform workers 55 regulation 60 rights 56, 57 skills 275 transformation 112 workplace tools 155, 157–9 digital disruption disruptive technologies and 3–4 human resource management and 4–6, 156–7 digital economy 124, 125, 128 defined 121 digital engagement model 157–9 digital governance 58–60 barriers to 63 design principles 62–5 of HRM operations 65–7 human resources and 60–62 workplace principles 64–5 digital immigrants 208, 210–11, 219, 220 digital tools fluency 217–18, 220 issues with the usage of 218 learning to use 217, 220 view on 215–17 goals and needs 214

online interviews 214–15 see also digital tools digitalization 110, 225 decision 169 and human adaptation 62 pandemic induced 168–9 talent management 166–8, 170, 172 algorithms 173 digitally disconnected workers 157–60 interviews with HR managers 160–61 digital natives 208, 210–11, 219, 220 online interview 214–15 view on digital tools 215–17 see also digital tools digital technologies (DTs) 3–4, 105–10, 112–15, 124, 126, 128 COVID-studies of 111–12 talent management 194 digital tools 207–10, 221 adaptivity of 220 on digital natives and immigrants 215–16 end-user demands 218–19 end-user of (see end-user, of digital tools) failing 216–17, 220 internet connections issues 216–17, 220 negative aspect of 216–17 positive aspect of 215–16 use of digi-generations 215–20 discrimination, in hiring processes 52, 53, 56, 67 disruptive debates 10–14 strategies 155 work environments 294 disruptive technologies 1, 156–7, 177, 224, 227, 295 and digital disruption 3–4 in HRM 180–81 MNEs 185–6 in talent management (TM) 180, 185–9 theorization of 180–81 distribution centres 160, 162 Dittes, S. 208 diversity management 275–6 Doberstein, C. 30 DTs see digital technologies (DTs) Duguid, P. 79 Dulebohn, J. H. 271 ecosystems, platforms 99

Index  303

Edwards, P. 264 Edwards, R. 23 Eginli, A. 208, 211, 220 e-leadership 105–7, 110–13 virtual team 106–13 electronic human resource management (e-HRM) 9, 136–7 applicant’s personality 146 automation of recruitment 145, 147 bibliometric analysis (see bibliometric analysis, in AI and HRM) recruiting and selection 143–7 training and performance evaluation 146 electronic leadership see e-leadership electronic performance monitoring (EPM) 273 Elkjaer, B. 78–82, 85, 87 El Sawy, O. A. 110 employee attributes 120, 122, 129 engagement 160, 161, 276, 277 employee-organization relationship 278 employment 11, 14 technology and 15–16 end-user, of digital tools 221 agencies 214–15 demands 218–19 failing digital tool 215–16 interactions negative experiences 216–17 positive experiences 215–16 engagement workforce (see workforce engagement) equal opportunities 275 Erofeeva, A. 212 ethics 51–2, 56–8 digital 60 in leadership research 113 Eubanks, B. 16 European Social Partners Framework Agreement on Digitalization 78 exclusive approach 178, 182, 186–8 see also inclusive approach Exclusive Tech Companion 187–9 Exclusive Tech Master 186–7, 189 experience 81–2, 87 exploratory interviews 160–61 external fit 94, 96, 98–9, 101 extrinsic motivation 25 face-to-face interactions 285, 287

fairness 54 metrics of 66–7 fault line 74, 86 feedback 124 algorithmic 27, 33 loop 61 systems 61 Festing, M. 182 Fleming, P. 257 Flexible Tech Companion 188, 189 Flexible Tech Master 187, 189 Ford, M. 15 framing talent 172–3 Frey, C. B. 10–11, 13, 15 Frost, A. C. 83 future of jobs, AI and 9–10, 15, 19 automation 10, 13, 16 disruptions 10–14 future of work 18, 169 Gagné, M. 23, 26 Galanaki, E. 147 Garavan, T. N. 275 Gardner, W. L. 107 Gegenhuber, T. 96 General Data Protection Regulation (GDPR) 53–4, 56, 68 article 25 of 56 article 88 of 76 generations digital (see digi-generations) Gibson, J. J. 240 gig economy 31–3 gig workers 96–100 Golden, T. D. 284 Gorman, B. 153 governance of AI 75 corporate 58–9, 64 defining 58–60 digital (see digital governance) motivations for 57–8 Grote, G. 255, 257–8 Habermas, J. 63 Haenlein, M. 225 hard skills 275 Harsch, K. 182 Hayes, L. 27 HCI see human-computer interaction (HCI) Heintzman, R. 153 hiring 43, 52–3, 146

304  Research handbook on human resource management and disruptive technologies

AI 42, 52–3, 55, 57, 68 assessment 200–201 outputs 199–200, 202 AI tool in (see AI hiring tool) algorithms 41–2, 44, 54, 195, 197–202 automation 62, 145–7 decision-making in 66, 196, 201–2 discrimination 52, 53, 56, 67 HR roles in 194–5, 197, 201–2 as gatekeeper 197–8 as guardian 199–200 as mediator 200–201 platform 41 Hoffman, G. 256 Holubová, B. 74, 86 Horn, R. E. 159 HRM see human resource management (HRM) HR managers, workforce engagement 154–7 disruptive strategy 155 disruptive technology 156–7 exploratory interviews 160–61 information design 159, 160 HR professionals 193, 196, 234 AI tools for 194–5 expert-based selection criteria 199 selecting candidates 197–8 selecting data 198 strategic roles 194–5, 197, 201–2 changing with AI 197–201 Hu, L. T. 291 human in the loop 29–30, 61–2 on the loop 62 out of the loop 62 human agency 209, 212 human-AI disputes 201 human bias 42–3 vs. algorithmic bias 66–7 human-cobot collaboration (see collaboration, human-cobot) human-computer interaction (HCI) 160 human-digital interaction 64–5 human resource management (HRM) 23–4, 39–40, 51, 54, 60, 114, 193, 194 algorithm 145 artificial intelligence (AI) in 136–7 bibliometric analysis (see bibliometric analysis, in AI and HRM) biases in developing AI for 40–43

and customer value 121, 125–9 decision-making 170–71 and digital disruption 4–6, 156–7 disruptive technologies in 180–81 functions 135 leadership and 114 policies 170 practices diversity management 275–6 new challenges 277–8 performance management 273 recruitment and talent management 273–4 sustainable management 276 training and development 274–5 principles for digital governance in 65–7 procedural and distributive justice 171 recruiting and selection 135–6 systems 94–5 in online labour platforms (OLPs) 96–101 research on mis-fit 99–100 talent management vs. 170–71 value co-creation 122–5, 129–30 see also talent management human resources (HR) and digital governance 60–62 governance 58 implications 271 skills 277 human rights 51–3 right of peaceful assembly and association 57 right to desirable work 55–7 right to education 55 right to equality 54–5 Huysman, M. 82 Huzzard, T. 85 hybrid decision-support tool 194 hybrid HRM system 97 hybrid work 272, 274, 276, 294 Hyman, R. 83, 84 hypothesis testing 291–2 Iacono, C. S. 180 Illeris, K. 81 inclusive approach 178–9 inclusive organizations 275 individuals acquisition 78 Industry 4.0 256–7, 264 informal learning environment 79

Index  305

information design 159, 160 technology 107, 108, 126, 170 informed consent 54 innovations, AI 18, 20 inquiry 80, 81, 87 institutional fit 94–9, 101 instrumental legitimacy 63 internal fit 94, 96, 97–8 internet 121, 125, 128 configuration services 126–7 networking service 127–8 solution services 125–6 interpersonal disputes 200 Intezari, A. 39 intra-organizational labour platforms (IOLPs) 238–40 case study of STRETCH affordances theory 244–9 data analysis 242–3 data collection 242 description 241–2 one-government mission 242, 246, 247 demand and supply 239, 249 material dimension 240–42, 245, 247–51 organizing tensions in 238–50 social dimension 240, 242, 245, 247–51 sociomateriality 241, 245 technology affordances and constraints 240–44, 248–9, 251 intrinsic motivation 25, 30–31 IOLPs see intra-organizational labour platforms (IOLPs) Isik, S. 208, 211, 220 Jackson, B. 85 Jacobides, M. G. 100 jobs automation 10, 13 characteristics 255, 257, 258, 264–5 human-cobot impact on 258–64 design 124, 257–8, 264 loss 257 non-automation 10 role changes 231–5 towards task 13 Kadir, B. A. 261 Kaplan, A. 225 Kayworth, T. R. 108

Kellogg, K. C. 23–4, 75 Kelly, E. 108 Kenda, R. 108, 110 Kesharwani, A. 208, 211 Keupp, M. M. 138 keywords analysis 142–3, 147 kiosks 155–7, 161–2 Kirby, J. 16, 17 knowledge acquisition 78, 79 and experience 81 Koroteev, D. 156, 157 Krakowski, S. 225 labelling datasets 43 labour market shortages 18, 20 labour productivity 15 Lamers, L. 29 Larjovuori, R. L. 110 Larson, L. 108 Lave, J. 79, 82 leadership 106–7 COVID-studies of 111–12 digital 105, 106, 110–13 electronic (see e-leadership) and HRM 114 organizational 251 research ethics 113 and technology 111 transformational 108, 109, 112 learning and development 245, 249, 251 organizational see organizational learning (OL) as participation 79–80 spaces 86 trade unions and 82–4 Lee, M. R. 113 left-over tasks 262, 265 legitimate decision-making 63 Leonardi, P. M. 209, 210, 212 Lévesque, C. 83 Levy, F. 17 Liao, C. 108 Lin, W. L. 114 Liversage, B. 114 living labs 57 Li, W. 108, 109 Lowe, K. B. 107 McAfee, A. 15, 16 McCarthy, P. 110

306  Research handbook on human resource management and disruptive technologies

Macey, W. H. 161 machine learning (ML) 40, 42, 43, 52, 54, 55, 136–7, 262 algorithms 193, 194, 196, 201 management control system 61 management fashion 9–10 March, J. G. 77, 78 Marler, J. H. 172, 194 Marshall, G. W. 284, 286, 292 Marson, B. 153 measurement collegial isolation 289–90 contextual work performance 289 relatedness 290 media agency case study 232, 234 Meijerink, J. 24, 33, 75 Mellahi, K. 178 mental workload 263 Miller, B. 15 mitigating bias in artificial intelligence 43–5 techniques 45–6 ML see machine learning (ML) MNEs see multinational enterprises (MNEs) Moeller, S. 125 Möhlmann, M. 26 Moore, P. 43, 257 Moore, S. 27 motivation 23–6 algorithmic management (AM) 28–33 characteristics of system 28–30 policies and practices 30–33 extrinsic 25 for governance 57–8 intrinsic 25, 30–31 motor-themes 144, 147 Mulki, J. P. 289 Müller-Jentsch, W. 82 multinational enterprises (MNEs) 177, 181 disruptive technologies 185–6 talent management (TM) 182, 185–6 Murnane, R. 17 Murray, G. 74, 75, 83 needs autonomy 26–7, 29 competence 26–9 psychological 285, 287, 293–5 relatedness 26, 27, 30 networking service, internet-enabled 127–8 Newman, D. T. 29 niche-themes 144, 147

Nilsson, N. J. 225 non-automation 10 nonsystemic interventions 67 NOPA+ model (network, openness, agility, participation, trust) 110, 113 Northcott, J. 75 O’Donovan, D. 147 OL see organizational learning (OL) Ollerenshaw, J. A. 213 OLPs see online labour platforms (OLPs) online interviews digi-generations 212–20 digital immigrants 214–15 digital native 214–15 end-user agencies 214–17 online labour platforms (OLPs) 95–101 organizational learning (OL) 74, 77–8, 87 collective participation 78–9 communities of practice (CoP) 79–80, 82 Elkjaer’s ‘third way’ of 80–82, 86 first and second way of 78–9 individuals 78, 80–81 knowledge acquisition 78, 79 orders 84 pragmatism 80–82 trade unions and 84–5 transactional approach 81 unit of learning 79 organizing tensions, in workplace 238–50 organizational capacity model 83 isolation 289 leadership 251 typology 186–9 Orlikowski, W. J. 180, 209 Osborne, M. A. 10–11, 13, 15 Östergren, K. 85 Ostroff, C. 94, 95 Paleyes, A 45 pandemic, and digitalization 168–9 paradoxes 264, 265 paradox theory 227–8 coexistence of people and AI 228–9 Parent-Rocheleau, X. 23, 24 Parker, S. K. 24, 26, 255, 257–8 Parry, E. 208, 211, 220 Pauleen, D. J. 39 pay stability 32

Index  307

unpredictability 32 pay-for-performance (PfP) 32, 33 performance appraisal 146 management 123, 129, 273 metrics 31 targets 31 personal data protection 53–4 Pessach, D. 194 Petry, T. 110 PfP see pay-for-performance (PfP) Pinsonneault, A. 287 platforms 97–8 digital labour 55, 238, 239 ecosystems 99 intra-organizational labour see intra-organizational labour platforms (IOLPs) labour 238, 239 online labour 96–101 Pollak, A. 263 Ponce Del Castillo, A. 76 Porfírio, J. A. 110 potential biases, in developing artificial intelligence human biases 42–3 statistical and computational bias 41–2 systematic bias 40–41 power resources model 83–4 pragmatic learning 124 pragmatism 80–82 privacy 53–4, 56, 65 procedural legitimacy 63 productivity management 30–31 monitoring 30 professional isolation 286, 292, 294, 295 roles 197–201 proxies, in algorithmic hiring 41 psychological needs 285, 287, 293–5 autonomy 26–7, 29 competence 26–9 relatedness 26, 27, 30 Purvanova, R. K. 108, 110 Pygmalion effect 66 qualitative research method 177, 181 Raisch, S. 225 Ramirez, P. 264 RBV see resource-based view (RBV)

recognition 154, 159 recruitment and selection 135, 136, 143, 145–7 and talent management 273–4 Redman, T. C. 168 regulation algorithmic management 76 artificial intelligence China 69 EU 68, 76 US 68–9 of work and employment 75–6 regulatory developments 68–9 sandbox 57 re-humanizing 232–3 relatedness 286, 293 feelings of 288 hypothesis testing 291–2 limitations and future research 294–5 measure of 290 mediating role of 285, 287–8 model characteristics 290 needs 26–7, 30 psychological needs 26, 27, 30 social isolation and 293 see also collegial isolation; contextual work performance remote work 272, 274, 276 contexts 293 in COVID-19 pandemic 272, 274, 284, 285, 287, 292 see also hybrid work resource-based view (RBV) 100 resource orchestration 125 mediating effect of 128 responsible science 23 review, leadership and technology 111 revitalization, trade unions (TUs) 83, 84 right to desirable work 55–7 to education 55 to equality 54–5 of peaceful assembly and association 57 protection 76 risk, of automation 10 Robey, D. 209, 210 Robinson, A. 257 robotics 11, 15 collaborative 256–7, 261–4 process automation 145 Roman, A. V. 109, 114

308  Research handbook on human resource management and disruptive technologies

Salas-Vallina, A. 112 Sarti, D. 110 Schneider, B. 161 Schon, D. A. 78, 84–5 Schouteten, R. 16 Schumpeter, J. A. 257 Schwartz, R. A 40 SEC model see six e-competence (SEC) model selection/exclusion bias 42–3 self-determination theory (SDT) 25–6, 32, 285, 287, 292–3 algorithmic management (AM) and 26–8 self-determined motivation 33 self-fulfilling prophecies 67 Senders, J. T. 262 Senge, P. M. 77 service-oriented HRM systems 121, 122 Sfard, A. 78 Sharpp, T. J. 109 Simon, H. A. 77, 78 six e-competence (SEC) model 109, 111, 113 skills building 19 cobot technology impact on 264 development 263–4 digital 275 hard 275 human resources (HR) 277 soft 275 Slee, M. 79, 84, 86 Smids, J. 258, 262, 264 Smith, A. 11 Smith, H. 84–5 social data 41 relationships 33 worlds 82, 86, 87 social isolation measure of 290 and relatedness 293 sociomaterial artifact 212 sociomateriality 241 soft skills 275 Soleimani, M. 39, 45 solution services, internet-enabled 125–6 Sosik, J. J. 107 Spagnoli, P. 111 staffing 122–3 stakeholders alignment 64

statistical bias 40–42 strategic disruption 157 strategy, in talent management 170, 172, 174 Strauss, A. 82 Strickland, L. H. 106 Strohmeier, S. 157 subjective decisions 174 surveillance and monitoring, of workers 30 sustainable HRM 276 systemic bias 40–41 interventions 67 talent 166, 168, 172–4 mobility 274 talent management (TM) 19, 177, 190 debates on 178–9 decision-making 170–74 defined 171 digitalization 166–8, 170, 172 algorithms 173 digital technology 194 disruptive technologies in 180, 185–9 exclusive approach to 178, 182, 186–8 framing the 173–4 humans view 174 inclusive approach to 178–9 judgments of value 171–2, 174 MNEs 182, 185–6 policies 170 recruitment and 273–4 strategy 170, 172, 174 subjective decisions 174 technology in 168, 185 vs. human resource management 170–71 workforce differentiation 171 Tansley, C. 172 tasks cognitive 262, 263 intersection 258 nonroutine manual 263 performance 286, 289 routine manual 262 variety of 261–2 technological adaptivity 209–13, 219–21 agency 212 artifacts 185, 187, 209–12 change 15 development 11, 17, 274, 277, 278 disruption 12

Index  309

innovation 11–12, 168 and labour productivity 15 technology 167 agency and 209–10 collaborative 256–7 decision-making in 169–70 disruptive see disruptive technologies and employment 15–16 end users 209–10, 212 ensemble view 181 epistemologies 180 information 170 proxy view of 181, 185, 187–8 in talent management (TM) 168, 185 theorization 180–81 tool view of 170, 180–81, 185, 188 use of HR professionals 195 technology affordances and constraints 240–44, 248–9, 251 Tekic, Z. 156, 157 tele-homeworking 283 COVID-19 pandemic 285, 287, 294, 295 thematic map analysis 143–7 theorization, of disruptive technologies 180–81 Tilvawala, K. 211 Torre, T. 110, 111 trade unions (TUs) 74, 77, 87 algorithmic management 85–6 capabilities 83–4 characteristics 83 collective representation 83 and learning 82–4 and organizational learning 84–5 revitalization 83, 84 role of 76, 82–3 traditional engagement strategies 157 training 123 and development 274–5 transactional approach 81 relationships 82 transformational leadership 108, 109, 112 transnational agreements 77 transparency, of algorithmic systems 28 Uhde, A. 30, 33 Ulrich, D. 271 unemployment 12, 18 rate 14 reporting bias in 42

Urwin, P. 208, 211, 220 Utterback, J. M. 156 value co-creation 121–2, 128–9 HRM 122–5, 129–30 internet in 121, 125, 128 configuration services 126–7 networking service 127–8 solution services 125–6 value-in-use creation see value co-creation value propositions 123, 125, 128 Van Gelder, M. 283 Van Veldhoven, M. 283 Van Wart, M. 109 Vanzo, A. 256 Vargas, T. L. 29 Venz, L. 112 Vermeulen, M. 109 virtual team leadership 106–13 voice assistants 64 Vrontis, D. 180 Waldkirch, M. 97 Wang, W. 207 Wattenhofer, R. 13 Waycott, J. 208, 211, 220 Weber, E. 110 Welfare, K. S. 262 wellbeing, employee 276, 277 Wenger, E. 79, 82 Whitehouse, R. 159 Wiblen, S. 194 Wilson, E. J. 110 Wingfield, N. 261 work design 257–8, 264 work disconnection, from time and place 271–3, 276, 277 worker’s autonomy 258, 260–61 workforce differentiation 171 diverse 275 HRM policies 170 polarization 12 workforce engagement 152–4 digitally disconnected workers 157–9 digital workplace tools for 155, 157–9 HR managers 154 disruptive strategy 155 disruptive technology 156–7 exploratory interviews 160–61 information design 159 work from home see remote work

310  Research handbook on human resource management and disruptive technologies

working conditions 55 working-for-data phenomenon 27, 30 work-life balance 198 boundaries 113, 276, 277 workplace autonomy in 64 isolation 283–4 organizing tensions in 238–50 principles digital governance 64–5 technology bans 152, 155

Wu, Z. X. 148 Yang, L. 294 Yerkes, M. A. 283 Yoo, Y. 108 Zaccaro, S. J. 108 Zeike, S. 110 Zimmermann, P. 108 Zoll, R. 84 Zuboff, S. 224