When do the EU AI Act obligations actually apply?
The prohibited practices in Article 5 have applied since February 2025 and the general-purpose AI model obligations since August 2025. The Digital Omnibus on AI moved the high-risk obligations, including EU database registration, from August 2026 to 2 December 2027, and the rules for AI in regulated products to 2 August 2028.
The regulation entered into force on 1 August 2024 and applies in stages.
| From | What applies |
|---|---|
| 2 February 2025 | Article 5 prohibited practices; AI literacy under Article 4 |
| 2 August 2025 | General-purpose AI model obligations |
| 2 December 2027 | High-risk obligations under Annex III, EU database registration included |
| 2 August 2028 | AI in products already covered by EU harmonisation legislation |
The last two dates moved. The Digital Omnibus on AI, Regulation (EU) 2026/1744, in force since 27 July 2026, postponed the high-risk deadline from August 2026 and the regulated-products deadline from August 2027. Those dates are settled law now, not a proposal.
The delay is less generous than it looks. Reaching a high-risk system’s compliance state requires classifying it, building the technical documentation, running the conformity assessment route that applies, and registering it — sequential work measured in months. And the parts already in force are in force: a prohibited practice was unlawful in February 2025 and remains so.
Classifying your systems is the step that has no reason to wait, because every other deadline depends on knowing which tier you are in.
Regulation references
- Regulation (EU) 2024/1689 (EU AI Act) · CELEX 32024R1689
Read next
A free Solo account runs a real classification on one of your own systems — no card, no sales call.
Run a free classification