For "high-risk" AI systems — those used in employment, education, biometric ID, critical infrastructure, law enforcement, migration, and similar contexts — the EU AI Act (Regulation 2024/1689) requires conformity assessments before placement on the market: risk management, data governance, technical documentation, transparency, human oversight, and accuracy/robustness testing. Post-market monitoring continues for the life of the system.
Most high-risk categories permit conformity assessment via the provider's self-declaration; biometrics require involvement of a notified body. Non-compliance attracts substantial administrative fines (up to 7% of global turnover for prohibited practices, lower tiers for other categories).