How EU rules govern AI in employment decisions
An OnLabor article explains how EU data-protection law and newer rules regulate AI used in workplace management. It focuses on the GDPR, the AI Act and the Platform Work Directive.
Key takeaways · 4
- 01
Check whether workplace AI performs employer functions: the AI Act generally classifies those systems as high-risk.
- 02
Plan for risk management, data quality, documentation, transparency, human oversight, accuracy and cybersecurity requirements.
- 03
Inform affected workers and their representatives when using AI for algorithmic management, and ensure the system’s role can be explained.
- 04
A nominal human check may not be enough: a 2023 court ruling treated a decision as automated when human participation had no qualitative impact on its outcome.
Three instruments form the framework
OnLabor describes EU regulation of AI-enabled managerial decisions as a combination of existing data-protection law and newer rules, with three instruments at its center.[1] The GDPR has applied since 2018 and provides the general framework for processing personal data.[1] It also gives individuals a right not to be subject to solely automated decisions that have significant effects.[1] The AI Act and Platform Work Directive were adopted in 2024, but the article says they were not yet fully applicable.[1] The AI Act can also reach companies outside the EU that want to place AI systems on the EU market.[1]
High-risk systems bring specific duties
The AI Act sorts AI uses into risk categories ranging from minimal to unacceptable; social scoring is one example of a prohibited use.[1] AI systems used to carry out employer functions are generally classified as high-risk, triggering requirements.[1] The listed obligations cover risk management, data quality, documentation, transparency, human oversight, accuracy and cybersecurity.[1] According to the article, oversight must be effective: systems must be capable of being overseen by people using appropriate human-machine interface tools.[1] Separate dates apply to high-risk obligations: December 2, 2027, for standalone systems and August 2, 2028, for AI integrated into regulated products, according to Legal500.[2]
Workers need information and meaningful review
The article says algorithms increasingly take on managerial roles in warehouses, factories, call centers and law firms, while AI can analyze granular data across multiple dimensions.[1] Employees may not know which inputs employers use or understand the grounds for a particular decision.[1] Employers must observe GDPR principles including lawfulness, transparency, purpose limitation and data minimization.[1] For algorithmic management, employers must inform workers’ representatives and affected workers that they will be subject to the system.[1] The deployer must give a clear and meaningful explanation of the system’s role in decision-making.[1] The Platform Work Directive provides rights to explanations, human review and rectification for people whose work is subject to a decision made or supported by an automated system.[1]
Cases and timelines put safeguards in focus
The article cites a 2023 Court of Justice ruling that a decision can count as automated even if a human participates, when that participation has no qualitative impact on the outcome; the decision was made when the automated credit score was generated.[1] It also gives algorithmic firing as an example of a decision incompatible with EU law.[1] In 2021, Italian data-protection authorities fined Foodinho and Deliveroo approximately €2.5 million over rider-scoring and work-allocation systems that lacked sufficient transparency or opportunities to challenge outcomes.[1] The Amsterdam Court of Appeal found in 2023 that Uber’s human review of driver deactivations was merely symbolic.[1] Member states must implement the Platform Work Directive by December 2, 2026, according to Lewis Silkin.[4]
For teams deploying workplace AI, the key question is not only what a system decides, but whether workers can understand and challenge its role. Build meaningful human review, clear explanations and data-protection checks into deployment planning rather than treating them as afterthoughts.
Why it matters
AI training for your team
kju.ai is built for teams: daily sessions, team analytics, and rollout support for your organization.
How this developed
15 November 2026
How EU rules govern AI in employment decisions
Sources
- Regulating AI in the Employment Relationship in the European Union ✦ OnLaboronlabor.org
- AI In HR: More Time Through The Digital Omnibus On AI – Compliance Remains A Priority | Legal 500legal500.com
- When does the AI Act go into effect? What is its timeline for implementation? | AI Act Service Deskai-act-service-desk.ec.europa.eu
- The EU Platform Work Directive: what international employers need to know | Lewis Silkinlewissilkin.com
