AI Act: high-risk systems in sensitive areas
High-risk AI systems listed in Annex III (employment decisions, access to essential services, critical infrastructure management, law enforcement and similar) are subject to the strict obligations from 2 December 2027. That date comes from the AI Omnibus that entered into force on 27 July 2026; the original date was August 2026. Deployers need a risk management system, human oversight, logging and a fundamental rights assessment where required.
Ticks are kept in this browser only. Sign in to the client portal to share them with your team.