EU lawmakers propose a broad workplace algorithmic management directive
The European Parliament’s Committee on Employment and Social Affairs has issued a draft report urging a dedicated directive on algorithmic management in the workplace, extending protections to solo self‑employed persons as well as workers. The initiative responds to gaps in existing EU and national frameworks—including the GDPR, the AI Act, the Information and Consultation Directive, and the Directive on Transparent and Predictable Working Conditions—and goes beyond the Platform Work Directive’s sector‑specific reach. It aims to address digital management tools that monitor, supervise, evaluate, or support decisions affecting working conditions, even where systems do not qualify as AI under the AI Act.
The draft defines algorithmic management broadly as automated systems that oversee activities or make or support decisions materially impacting work assignments, pay, safety and health, working time, training, promotion, and contractual status. It would require written transparency to workers, representatives, and solo self‑employed individuals, covering the purpose of systems, categories of data collected (including behavior and performance data), types of monitored actions, and the scope of any automated decision‑making. Consultation with employee representatives would be mandatory prior to deployment or significant updates, addressing objectives, affected processes and workers, workload and scheduling effects, health and safety implications, data categories, safeguards, human oversight, and training.
The proposal would prohibit processing certain sensitive data (e.g., emotional state, private conversations, off‑duty behavior) and mandate effective human oversight for critical decisions such as hiring, termination, and pay changes. It would also introduce health and safety assessments tailored to algorithmic tools, focusing on accident risks, psychosocial and ergonomic impacts, and undue pressure, with national labor inspectorates empowered to supervise compliance—including non‑discrimination in use. These measures reflect the growing role of advanced models in people management and seek to ensure accountability regardless of whether systems qualify as AI.
Procedurally, the Committee must approve the report before a plenary vote expected in November. If adopted, the European Commission would decide whether to initiate legislation, triggering the ordinary legislative process involving the Parliament and Council. If enacted and transposed into national law, organizations would face audits of existing systems, strengthened grievance mechanisms, and policy overhauls across data protection, performance management, and worker rights (potentially including a right to disconnect). Targeted training for managers and HR would become essential. Given the breadth of the definition and potential operational impact, affected organizations should closely track legislative developments.
