Do algorithms play fair? Analysing the perceived fairness of HR-decisions made by algorithms and their impacts on gig-workers

• Published in: International journal of human resource management Vol. 36; no. 2; pp. 235 - 274
• Main Authors: Jabagi, Nura, Croteau, Anne-Marie, Audebrand, Luc K., Marsan, Josianne
• Format: Journal Article
• Language: English
• Published: London Routledge 19.01.2025; Taylor & Francis LLC
• Subjects: Algorithmic management; algorithmic perceptions; Algorithms; digital labour platforms; Fairness; Gig economy; HR algorithmic decisions; Human resources; Job satisfaction; Organizational justice; Perceptions; Performance evaluation; Transparency; Workers
• ISSN: 0958-5192; 1466-4399
• DOI: 10.1080/09585192.2024.2441448

On digital labour platforms, algorithms execute a wide range of human resource (HR) decisions including work allocation and performance evaluation. Despite their growing use, our understanding of how people perceive such algorithms, particularly in terms of fairness, is less developed. Using Organisational Justice Theory, we explore how workers perceive the fairness of HR-decisions made by algorithms and how those perceptions impact job satisfaction and perceived organisational support (POS). Results from a survey of 435 Uber drivers indicate that perceptions of algorithmic fairness - and their formation - differ based on the type of HR-decision enacted by an algorithm and whether those decisions are considered to require mechanical or human skills. Results also demonstrate positive significant relationships between perceived algorithmic fairness, POS, and job satisfaction. This study answers calls to investigate perceptions of algorithmic fairness across different HR-decisions and their impacts in real-world settings. Our results suggest that algorithms play an important role in shaping platform-workers' experiences and attitudes as both technological artefacts and social agents of the organisation. Recommendations for improving the perceived fairness of algorithms for HR-decisions by focusing on transparency and high impact/value fairness indicators are offered.