If you hire or employ people in the EU, the EU AI Act sets rules for AI in employment — some already in force, others from December 2027.
Already in force: banned practices
Since 2 February 2025, the Act prohibits AI systems that infer the emotions of a person in the workplace (except for medical or safety reasons). Tools that claim to read engagement, stress or honesty from faces or voices at work fall here.
Already in force: AI literacy
Article 4 asks providers and deployers of AI to take measures for a sufficient level of AI literacy among staff who use AI systems. The 2026 Digital Omnibus changes how this duty is worded; check the final text with counsel. Either way, AI training is now part of compliance, not only a nice-to-have.
From 2 December 2027: high-risk employment AI
Annex III, point 4 lists as high-risk:
- AI to recruit or select people, including targeting job ads, analysing and filtering applications and evaluating candidates;
- AI to make decisions on terms of work, promotion and termination, to allocate tasks based on behavior or traits, or to monitor and evaluate performance and behavior.
The original date was 2 August 2026. Under the Digital Omnibus agreement announced by the European Commission on 7 May 2026, high-risk rules for these areas apply from 2 December 2027.
What employers (deployers) must do
Among other duties, deployers of high-risk AI must use it according to the provider's instructions, assign human oversight to competent people, keep logs, monitor operation, and — under Article 26(7) — inform workers' representatives and the affected workers before putting a high-risk AI system into use at the workplace.
Practical takeaway
Use the extra time. Inventory employment AI now, remove any emotion-recognition features, ask vendors how they will meet provider obligations, and design worker information and oversight into rollouts — works councils in many EU countries will expect to be consulted anyway.