Regulatory clock
What applies and when. This page updates from a feed; check back as deadlines approach.
| Deadline | Obligation | Status |
|---|---|---|
| Prohibited practices (Article 5) and AI literacy duties | In force | |
| General-purpose AI model obligations | In force | |
| Article 50 transparency: people must be told they are interacting with AI; synthetic content marked (marking grace period to 2 Dec 2026 for systems already on the market) | In force | |
| Annex III high-risk AI systems: risk management, data governance, documentation, record-keeping, human oversight, accuracy and robustness | Upcoming (deferred from 2 Aug 2026) | |
| Annex I high-risk AI systems (AI as a safety component of regulated products) | Upcoming (deferred from 2 Aug 2027) |
Note: The high-risk dates were deferred by the 2026 omnibus (political agreement 6 May 2026, confirmed by Member States 13 May 2026); the transparency date was not. Dates are checked against the published text and updated here as guidance evolves.
See AI assurance for readiness assessment and gap analysis.