Umeå universitets logga

The increasingly pervasive use of digital technologies, especially algorithms and artificial intelligence (AI) in the workplace, has raised growing concerns about the need to identify and regulate their effect on occupational safety and health (OSH). The European Union (EU)‘s Artificial Intelligence Act (AI Act) represents a significant regulatory framework aimed at overseeing the development and deployment of AI technologies within the EU. This paper critically examines the effectiveness of this legislation in managing occupational safety and health (OSH) risks.

While the AI Act classifies many algorithmic management (AM) systems as“high-risk” and imposes ex ante obligations on providers to assess and mitigate safety risks, its product safety-oriented approach is limited in addressing the context-specific, dynamic, and cumulative nature of workplace harms. Drawing on a doctrinal analysis of the AI Act (Regulation (EU) 2024/1689) and the EU OSH Framework Directive (89/391/EEC), this paper compares the AI Act’s risk management requirements with the ongoing, participatory, and preventive duties embedded in OSH regulation. Findings highlight critical regulatory gaps, particularly limited obligations of employers (“deployers”) under the AI Act and the challenges of addressing psychosocial risks through design measures alone. The article argues for a regulatory approach that aligns the AI Act with existing OSH frameworks and calls for greater guidance, oversight,and worker involvement in the deployment of algorithmic systems. Only by recognising AM as both a technical product and a socio-organizational practice can regulation effectively safeguard worker health, safety, and dignity.