EU AI Act timeline

In phases. Article 4 AI literacy and Article 5 prohibited practices became enforceable on 2 February 2025 and apply today. General-purpose AI provider obligations under Articles 50-55 became enforceable on 2 August 2025. The high-risk Annex III standalone obligations apply from 2 December 2027 (delayed from 2 August 2026 by the Digital Omnibus deal of 7 May 2026). High-risk AI embedded in regulated products under Annex I applies from 2 August 2028. The Article 99 penalty regime is enforceable today for any violation in scope.

Yes, for the high-risk obligations. On 7 May 2026 the EU Council and Parliament reached a provisional agreement (the Digital Omnibus on AI) that delayed the high-risk Annex III standalone deadline from 2 August 2026 to 2 December 2027 (a 16-month shift) and the Annex I embedded high-risk deadline to 2 August 2028. The rest of the regulation timeline is unchanged: Article 4 AI literacy, Article 5 prohibited practices, GPAI provider obligations under Articles 50-55, and the Article 99 penalty regime all continue on their original schedule.

Two of the most important obligations of the EU AI Act became enforceable. Article 4 AI literacy requires every organisation deploying AI in the EU to ensure staff has role-proportionate literacy. Article 5 prohibits specific AI practices, including social scoring by public authorities, manipulative AI, untargeted facial-image scraping, emotion recognition in workplaces and schools, real-time biometric mass surveillance, predictive policing based solely on profiling, and biometric categorisation for sensitive characteristics. Both obligations are live and enforceable today with the full Article 99 penalty regime applying.

Annex III standalone high-risk AI systems come under the full Article 8-49 obligation stack. The eight Annex III categories are: biometrics, critical infrastructure, education and vocational training, employment and worker management, access to essential services (credit, benefits, healthcare, emergency dispatch), law enforcement, migration / asylum / border control, administration of justice and democratic processes. Deployers face Article 26 obligations (human oversight, transparency, incident reporting) and Article 27 fundamental rights impact assessments. Providers face Article 8-22 conformity assessment obligations. National market surveillance authorities begin active supervision on this date.

High-risk AI embedded in regulated products under Annex I becomes subject to the AI Act regime on top of existing sectoral conformity assessment. Annex I covers AI used as safety components of lifts, toys, medical devices, in-vitro diagnostic devices, machinery, vehicles, marine equipment, civil aviation security, and similar regulated product categories. The AI Act layers on top of the existing sectoral frameworks (MDR, IVDR, Machinery Regulation 2023/1230, etc.); providers must demonstrate conformity with both regimes simultaneously.

The provisional agreement of 7 May 2026 needs formal adoption by both the European Parliament (plenary vote) and the Council of the EU before entering into force. Both adoptions are scheduled for the second half of 2026 on standard EU legislative timelines, with publication in the Official Journal expected by late 2026 or early 2027. The new deadlines (December 2027 and August 2028) are conditional on this adoption process completing. Organisations should plan around the new dates but watch for any procedural disruption in the formal adoption phase.

No. The timeline tool surfaces the dates set out in Article 113 of Regulation (EU) 2024/1689 as amended by the Digital Omnibus on AI provisional agreement of 7 May 2026. For binding interpretation of how those dates apply to a specific AI system, organisation, or use case, consult a qualified EU technology lawyer or your national market surveillance authority. Member State implementations may add specific national procedures on top of the EU framework.