CSOAI - enforcement timeline
The EU AI Act, date by date
From entry into force to the last high-risk deadline - exactly what applies when, including the Digital Omnibus deferral. The next cliff is highlighted.
- 1 Aug 2024In effectEntry into force
The EU AI Act becomes law; the enforcement clock starts.
- 2 Feb 2025In effectProhibited practices banned
Unacceptable-risk AI (social scoring, manipulative systems, untargeted facial scraping) is outlawed. AI-literacy duties begin.
- 2 Aug 2025In effectGPAI rules + governance start
General-purpose AI model obligations begin for new models; the AI Office and penalty framework come online.
- 2 Aug 2026In effectTransparency + penalties enforceable
Article 50 transparency duties apply, alongside the full penalty/market-surveillance regime and AI Office GPAI enforcement powers. The headline 2026 cliff.
- 2 Dec 2026Next cliffLegacy marking grace ends + new Art 5 bans
Machine-readable marking (Art 50(2)) becomes mandatory for generative systems placed on the market before 2 Aug 2026; new Article 5 prohibitions on NCII/CSAM generation take effect.
- 2 Aug 2027UpcomingLegacy GPAI compliance + sandboxes
GPAI models placed on the market before 2 Aug 2025 must reach compliance; the national regulatory-sandbox obligation applies.
- 2 Dec 2027UpcomingAnnex III high-risk obligations
Full obligations for stand-alone high-risk systems (Annex III) - deferred from Aug 2026 by the Digital Omnibus (Reg (EU) 2026/1744, in force 27 July 2026).
- 2 Aug 2028UpcomingHigh-risk in regulated products
High-risk AI that are safety components of products already regulated under EU law (Annex I) must comply - deferred by the Digital Omnibus.
