Annex III, point 4 covers AI intended for recruitment, candidate selection, and decisions affecting employment relationships.
AI that reads, scores, or filters job applications, including keyword matching, skills extraction, and automated shortlisting.
Psychometric tools, personality assessments, or competency tests that use AI to evaluate candidates and influence hiring.
Systems that analyse facial expressions, tone of voice, or body language during video interviews.
AI used to recommend promotions, flag performance issues, or influence retention and termination decisions.
Operational and administrative tools that don't directly influence individual hiring outcomes are generally lower risk.
Even if your system isn't high-risk, Article 50 transparency may still apply. Run the free classifier to find out.
Each must be in place before December 2, 2027. Non-compliance risks fines up to €15 million or 3% of global turnover.
Employment AI is widely expected to be among the first systems regulators scrutinise. Article 10 requires detailed documentation of training data, bias testing across protected characteristics (gender, age, ethnicity, disability), and mitigation measures. If your AI scores candidates, you need to show it doesn't discriminate, and document how you tested for it.
Read: EU AI Act for HR ToolsGDPR Article 22 already restricts automated decision-making in employment. If you have DPIAs and lawful-basis documentation for your AI hiring tools, that partially covers AI Act Articles 9 and 10. But the AI Act adds what GDPR doesn't: model accuracy records, conformity assessment, bias testing documentation, and continuous post-market monitoring.
See the full GDPR overlap mappingRecruitment AI will be among the first systems regulators examine. The classifier is free; classify your system and start generating the documentation you need.
Classify your AI system