AI Act Timeline
Complete timeline for the entry into force of Regulation (EU) 2024/1689 on artificial intelligence. Four key deadlines between February 2025 and August 2027.
Prohibited practices and AI literacy
Entry into force of the prohibitions under Article 5 and the AI literacy obligation under Article 4. Organizations must ensure that their personnel have a sufficient level of AI understanding.
Applicable obligations:
- Prohibition of social scoring systems (Art. 5(1)(c))
- Prohibition of subliminal manipulation and exploitation of vulnerabilities (Art. 5(1)(a)(b))
- Restrictions on real-time remote biometric identification (Art. 5(1)(d))
- Prohibition of biometric categorization based on sensitive data (Art. 5(1)(g))
- AI literacy obligation for all relevant personnel (Art. 4)
GPAI rules and governance
Application of obligations for general-purpose AI models (Chapter V) and establishment of European governance. The European AI Office oversees GPAI model providers' compliance.
Applicable obligations:
- Technical documentation for GPAI models (Art. 53(1)(a))
- Copyright compliance policy (Art. 53(1)(c))
- Publication of training data summary (Art. 53(1)(d))
- Enhanced obligations for systemic risk models (Art. 55)
- Designation of national competent authorities (Art. 70)
High-risk (Annex III) + Transparency (Art. 50)
Application of all requirements for high-risk AI systems under Annex III and transparency obligations under Article 50. This deadline covers the majority of the Regulation's substantive obligations for providers, deployers, importers and distributors.
The omnibus proposal (not yet adopted) would split this into two separate deadlines: transparency in November 2026 and high-risk in December 2027. These postponements will only take effect upon formal adoption.
Applicable obligations:
- Risk management system (Art. 9)
- Data governance (Art. 10)
- Technical documentation compliant with Annex IV (Art. 11)
- Human oversight and transparency (Art. 13, 14)
- Information for persons interacting with an AI system (Art. 50.1)
- Synthetic content watermarking (Art. 50.2)
- Deepfake content disclosure (Art. 50.4)
Systems integrated in regulated products (Annex I)
Final deadline of the application timeline, concerning high-risk AI systems that are safety components of products covered by Union harmonization legislation listed in Annex I (machinery, toys, medical devices, etc.).
The omnibus proposal (not yet adopted) would postpone this to 2 August 2028. This postponement will only take effect upon formal adoption.
Applicable obligations:
- Compliance of AI systems as safety components of Annex I products
- Integration into existing sectoral conformity assessment procedures
- Technical documentation integrated into product documentation
- CE marking covering both the product and the AI component
Frequently asked questions about the AI Act timeline
Find answers to the most common questions about the deadlines and phased implementation of the European AI Regulation.
Regulation (EU) 2024/1689 on artificial intelligence (AI Act) was published in the Official Journal of the European Union on 12 July 2024 and entered into force on 1 August 2024. However, its provisions apply progressively, with deadlines staggered between February 2025 and August 2027 depending on the type of obligation and the risk level of the AI system concerned.
The AI Act provides four main deadlines: 2 February 2025 for prohibited practices and the AI literacy obligation (Article 4), 2 August 2025 for rules applicable to GPAI models, 2 August 2026 for transparency obligations (Article 50) and Annex III high-risk systems, and 2 August 2027 for AI systems embedded in regulated products under Annex I.
Since 2 February 2025, two categories of obligations have been applicable: the prohibition of unacceptable-risk AI practices (Article 5), such as subliminal manipulation or social scoring, and the AI literacy obligation (Article 4), requiring all providers and deployers to train their staff. Since 2 August 2025, the rules for general-purpose AI (GPAI) models are also in effect.
The omnibus proposal is a legislative text put forward by the European Commission aiming to simplify and postpone certain AI Act deadlines. It notably proposes postponing transparency obligations (Article 50) to November 2026, Annex III high-risk systems to December 2027, and Annex I regulated products to August 2028. Until this proposal is formally adopted by the Parliament and Council, the official regulation dates remain applicable.
Under the current regulation, high-risk AI systems listed in Annex III (biometrics, critical infrastructure, employment, education, etc.) must be compliant by 2 August 2026. Systems embedded in regulated products under Annex I (medical devices, machinery, etc.) have until 2 August 2027. The omnibus proposal, if adopted, could postpone these deadlines to December 2027 and August 2028 respectively.
Non-compliance with obligations by the application date exposes organisations to significant financial penalties: up to EUR 35 million or 7% of worldwide annual turnover for prohibited practices, up to EUR 15 million or 3% for other obligations. National market surveillance authorities are responsible for enforcement and may conduct inspections, request documentation and order the withdrawal of non-compliant systems from the market.
Yes, the AI Act also applies to AI systems already placed on the market or put into service before the application dates, provided they undergo significant modifications after the regulation enters into force. High-risk AI systems already deployed by public authorities must be brought into compliance within the prescribed deadlines. Providers and deployers should therefore anticipate and plan their compliance efforts ahead of each deadline.
To prepare, it is recommended to: (1) carry out an inventory of all your AI systems and determine their risk level, (2) identify the obligations applicable to your role (provider, deployer, importer), (3) establish the required technical documentation and risk management system, (4) train your staff on AI literacy, and (5) plan the necessary conformity assessments. Our free diagnostic tool helps you quickly identify your obligations and the deadlines that apply to you.
Prepare before the next deadline
Free diagnostic in under 3 minutes. Identify the obligations applicable to your AI system and start your compliance journey.