Skip to content

EU AI Act in HR & Recruiting

AI in the application process, for promotion decisions or performance evaluations – Annex III No. 4 makes almost every HR AI a high-risk system. With particular requirements for non-discrimination.

Updated: March 202610 min read

HR as a core area of the EU AI Act

Annex III No. 4 of Regulation (EU) 2024/1689 explicitly lists AI systems in employment, workforce management and access to self-employment as high-risk. This includes:

  • AI systems for recruitment and selection of applicants (screening, ranking, pre-selection)
  • AI for promotion decisions and task allocation
  • AI for performance monitoring and evaluation of employees
  • AI for termination decisions

Typical high-risk scenarios in HR

ApplicationHigh-risk?Justification
CV screening / applicant rankingYesAnnex III No. 4 – recruitment
Video interview analysisYesAnnex III No. 4 + No. 1 (biometrics)
Employee performance scoringYesAnnex III No. 4 – performance evaluation
Promotion recommendationsYesAnnex III No. 4 – promotion
Emotion recognition in the workplaceProhibitedArt. 5 – Prohibited practices
Scheduling / shift planningNoNo impact on fundamental rights
Chatbot for HR FAQNoArt. 50 transparency obligation suffices

Particular discrimination risks

HR AI carries particularly high discrimination risks. The EU AI Act therefore requires comprehensive measures:

  • Data governance (Art. 10) – Training data must be checked for bias. Historical hiring data can contain discriminatory patterns.
  • Accuracy and robustness (Art. 15) – Performance metrics must be reported separately for different demographic groups.
  • Human oversight (Art. 14) – AI recommendations must not be implemented automatically. Human-in-the-loop is mandatory.

In parallel, the EU anti-discrimination directives apply: Directive 2000/43/EC (Racial Equality Directive), Directive 2000/78/EC (Employment Equality Framework Directive) and in Germany the AGG (General Equal Treatment Act).

FRIA obligation in HR

Under Art. 27 EU AI Act, deployers of HR AI systems under Annex III No. 4 must conduct a Fundamental Rights Impact Assessment (FRIA). Key focus areas:

  • Non-discrimination (Art. 21 EU-GRC) – Risks related to gender, age, origin, disability
  • Right to work (Art. 15 EU-GRC) – Access to employment
  • Fair working conditions (Art. 31 EU-GRC) – In performance monitoring
  • Data protection (Art. 8 EU-GRC) – Processing of applicant data

Special case: Works council and co-determination

In Germany, Austria and other EU countries, works councils have a co-determination right when introducing technical monitoring systems (§ 87(1)(6) BetrVG in Germany). AI systems in HR almost always trigger this co-determination requirement.

Tip: The EU AI Act documentation (especially the Technical Documentation and Transparency Notice) simultaneously serves as a basis for the works council consultation.

Documentation obligations for HR AI

  • Technical Documentation – For providers of HR AI tools
  • Transparency notice – Applicants and employees must be informed about the use of AI
  • FRIA – For deployers (HR departments using an external tool)
  • Bias documentation – Evidence of fairness testing for different demographic groups

Next steps

  1. Inventory – What AI tools does your HR department use?
  2. Risk checkCheck for free whether a high-risk system is present
  3. Involve works council – Inform early and provide documentation
  4. Create documentationGenerate compliance drafts with HR-specific context

Generate compliance drafts for HR AI now →

Compliance drafts

FRIA, Technical Documentation and Transparency Notice – AI-generated drafts in minutes.

Generate drafts

Ready for your EU AI Act Compliance?

AI-generated compliance documents as a solid working basis for you or your lawyers.

EU AI Act in HR & Recruiting | AIvunera