Compliance timeline
for the EU AI Act
Browse the hubCompliance timeline
Key dates and deadlines
Every confirmed deadline from entry into force through 2030, and the obligations attached to each.
Entry into force
The Act is published and in force. The transition clock starts; no obligations apply yet.
The Act became law on 1 August 2024 following publication in the Official Journal on 12 July 2024. Entry into force did not trigger compliance obligations, it started the transition clocks.
Organisations should have used this period to inventory their AI systems, identify which risk tier applies, and assign internal ownership for the compliance programme.
Who this affects: All organisations placing AI systems on the EU market or putting them into service in the EU.
Prohibited practices and AI literacy
Article 5 bans become enforceable with no transition period. Article 4 AI-literacy duties apply.
Six months after entry into force, the eight prohibited AI practices under Article 5 became enforceable. These are absolute bans, no derogation, no transition period. Systems falling within these categories had to be withdrawn.
The prohibited categories include:
- Subliminal, manipulative, or deceptive techniques that distort behaviour
GPAI model obligations
Chapter V obligations apply to providers of general-purpose AI models.
Chapter V obligations apply to providers of general-purpose AI models, models trained on broad datasets to perform a wide range of tasks, including large language models.
All GPAI model providers must:
- Maintain technical documentation
Article 50 transparency and deployer obligations
Chatbot disclosure, synthetic-content labelling and most deployer transparency duties apply.
Several obligations come into force on 2 August 2026 regardless of the Omnibus outcome:
- Chatbot disclosure (Article 50(1)): users must be told when they are interacting with an AI, unless obvious from context.
Deferred synthetic-content labelling and new prohibitions
Watermarking for pre-market systems; new Article 5 bans take effect.
Under the Omnibus provisional agreement, the Article 50(2) machine-readable labelling obligation for AI-generated content systems placed on the market before 2 August 2026 is deferred to this date.
Two new Article 5 prohibitions also apply from 2 December 2026: AI used to generate non-consensual intimate imagery, and AI used to generate child sexual abuse material.
National regulatory sandboxes
Each Member State must have at least one AI regulatory sandbox in place.
Member States must have established at least one national AI regulatory sandbox, a supervised environment where providers can develop and test systems before market, deferred from 2 August 2026 under the Omnibus.
Annex III high-risk systems (deferred)
Full obligations for stand-alone Annex III high-risk systems.
Under the original Act, the full suite of high-risk obligations was due on 2 August 2026. The Digital Omnibus provisional agreement defers this to 2 December 2027 for stand-alone Annex III systems, defined by use case:
- Biometric identification and categorisation
Annex I high-risk systems (deferred)
AI embedded in products already regulated under EU product-safety law.
AI systems embedded in products covered by existing EU product-safety legislation face an additional year. These are AI components inside physical products subject to prior CE-marking requirements:
- Medical devices (MDR/IVDR)
Public-authority legacy systems
High-risk AI used by public authorities before Aug 2026 must fully comply.
High-risk AI systems already in service by public authorities before 2 August 2026 have the longest transition and must reach full compliance by 2 August 2030.
This applies to Member State government agencies, public bodies, and public utilities deploying AI systems that predate the high-risk threshold date.
This is general information, not legal advice. Please consult your legal or compliance team to confirm your organisation's specific obligations.
Plan against the right dates
Deeploy maps every AI system in your inventory to the deadlines that apply to it, and updates the plan when the rules move.