// ai policy & safety
A sourced timeline of the EU AI Act from the 2021 proposal to the 2030 transition deadlines. It includes the 2026 Digital Omnibus that moved the high-risk rules to December 2027 and August 2028.
Researched and fact-checked by AI, with no human review. How we verify
25 entries from 24 sources. Checked for news weekly; last checked
This tracker lists milestones of the EU Artificial Intelligence Act, Regulation (EU) 2024/1689. It covers:
It is current as of October 6, 2026.
Entries cover events that changed the legal text, set or triggered an application date, or produced major final guidance on compliance. The guidance listed is a selection. Entries cite EU institutions' own pages or the Official Journal. Future dates are marked scheduled and follow the consolidated text of July 27, 2026. Under that text, obligations for high-risk AI systems start on December 2, 2027 and August 2, 2028. Those dates replace August 2, 2026 and August 2, 2027. The Commission describes high-risk systems as uses that can pose serious risks to health, safety or fundamental rights. Annex III lists stand-alone high-risk uses such as hiring or border control. Annex I covers AI built into regulated products.
Drafts and consultations are left out. That includes the draft guidelines on classifying high-risk systems published on May 19, 2026. Also left out are committee votes, national implementing laws, standards work, enforcement cases and commentary. The Commission keeps its own overview on its AI Act page.
Targets set by the organizations named. Dates like these often move.
Scheduled
Article 111, as consolidated on July 27, 2026, sets two 2030 deadlines for systems already in use. High-risk AI intended for public authorities has until August 2, 2030. AI components of large-scale EU IT systems in Annex X placed on the market before August 2, 2027 have until December 31, 2030.
Source: EUR-Lex - 02024R1689-20260727 - EN - EUR-Lex, Publications Office of the European Union (EUR-Lex)
Scheduled
Rules for high-risk AI embedded in physical products covered by Annex I, such as toys and lifts, start to apply on this date, according to the Commission. The Digital Omnibus on AI set the date. The Commission says the Omnibus entered into force on July 27, 2026.
Scheduled
Sections 1 to 3 of Chapter III apply from this date to AI systems classified as high-risk under Article 6(2) and Annex III. Regulation (EU) 2026/1744 of July 8, 2026 set the date in place of August 2, 2026. Its recitals cite late standards and delays in setting up national authorities.
Source: Regulation - EU - 2026/1744 - EN - EUR-Lex, Publications Office of the European Union (EUR-Lex)
Scheduled
Article 57, as amended in 2026, requires each member state to have at least one national AI regulatory sandbox, a supervised testing environment, operational by this date. Article 111 gives providers of general-purpose AI models placed on the market before August 2, 2025 the same deadline to comply.
Source: EUR-Lex - 02024R1689-20260727 - EN - EUR-Lex, Publications Office of the European Union (EUR-Lex)
Scheduled
Two bans added by the Digital Omnibus on AI start to apply. They cover AI systems that generate non-consensual sexual deepfakes or child sexual abuse material. Providers of generative AI systems on the market before August 2, 2026 must comply with the Article 50(2) marking rule by this date.
Source: Timeline for the Implementation of the EU AI Act, European Commission (AI Act Service Desk)
Article 50 transparency rules started to apply. Chatbots must tell users they are AI. Deepfakes must be labeled, and AI-generated content needs machine-readable marking. The Commission said on July 31, 2026 that its AI Office and national authorities would start enforcing the Act on this date.
The Council adopted the amending regulation on June 29, 2026. It was signed on July 8 and published in the Official Journal on July 24, 2026. It entered into force on the third day after publication. The regulation moved the high-risk dates to December 2, 2027 and August 2, 2028.
Source: Regulation - EU - 2026/1744 - EN - EUR-Lex, Publications Office of the European Union (EUR-Lex)
The Commission published guidelines on how far the Article 50 transparency obligations reach for providers and deployers of AI systems. It says it adopted them as practical guidance for companies and national authorities. The obligations started to apply on August 2, 2026.
Parliament adopted its position by 423 votes to 57, with 174 abstentions. The text applies high-risk obligations from December 2, 2027 and August 2, 2028. It bans AI that generates sexual imagery of identifiable people without consent. It also requires national sandboxes by August 2, 2027.
The final voluntary code on marking and labeling AI-generated content was published on the day of its closing plenary. It covers Article 50(2), (4) and (5). The Commission says about 190 companies and organizations had signed it by the end of July 2026.
Source: Code of Practice on Transparency of AI-generated Content, European Commission
According to the Commission, the deal fixed the high-risk timetable. It set December 2, 2027 for uses in areas including biometrics, critical infrastructure, education, employment and border control. It set August 2, 2028 for AI integrated into products such as lifts or toys.
Proposal COM(2025) 836 tied the high-risk rules to a Commission decision that compliance support was available. They would have started 6 months (Annex III) or 12 months (Annex I) after that decision, and no later than December 2, 2027 and August 2, 2028. It also proposed an Article 50(2) grace period to February 2, 2027.
Source: EUR-Lex - 52025PC0836 - EN - EUR-Lex, Publications Office of the European Union (EUR-Lex)
Obligations for providers of general-purpose AI models started to apply. The Commission's timeline also gives this date for member states to designate national competent authorities and adopt penalty laws. It gives the same date for the AI Board, Scientific Panel and Advisory Forum to be set up.
Source: Timeline for the Implementation of the EU AI Act, European Commission (AI Act Service Desk)
The Commission published guidelines on the scope of the obligations for providers of general-purpose AI models. That was about two weeks before those obligations started to apply on August 2, 2025. The Commission's page offers the text in all 24 official EU languages.
Independent experts prepared the voluntary code. It has three chapters: Transparency, Copyright, and Safety and Security. The Commission says that it and the AI Board have since confirmed the code as an adequate voluntary tool for model providers to show compliance with the Act.
Source: The General-Purpose AI Code of Practice, European Commission
The guidelines give the Commission's interpretation of the Act's bans. They include legal explanations and practical examples on practices such as harmful manipulation, social scoring and real-time remote biometric identification. The guidelines are non-binding.
The first provisions of the Act started to apply, the Commission said. They cover the definition of an AI system and AI literacy requirements. They also ban a small number of AI uses that the Act treats as posing unacceptable risk.
Source: First rules of the Artificial Intelligence Act are now applicable, European Commission
The Act entered into force on this date. The Commission said the general-purpose AI provisions would apply 12 months later. It also said it expected to finalize a code of practice for those models by April 2025.
The Act was signed on June 13, 2024. It was published with entry into force set for the twentieth day after publication. Article 113 set general application for August 2, 2026. It set August 2, 2027 for Article 6(1) high-risk systems. Some chapters had earlier dates in February and August 2025.
Source: Regulation - EU - 2024/1689 - EN - EUR-Lex, Publications Office of the European Union (EUR-Lex)
The Council adopted the Act after Parliament's first reading, according to Parliament's Legislative Observatory. It records the final act as signed on June 13, 2024 and published in the Official Journal on July 12, 2024.
Source: Procedure File: 2021/0106(COD), European Parliament (Legislative Observatory)
MEPs endorsed the regulation by 523 votes in favor, 46 against and 49 abstentions. It had been agreed with member states in December 2023. The text bans uses such as social scoring and untargeted scraping of facial images to build facial recognition databases.
Source: Artificial Intelligence Act: MEPs adopt landmark law, European Parliament
Negotiators agreed on a provisional text. It included bans on social scoring and on emotion recognition in workplaces and educational institutions. Parliament's press release is time-stamped 00:04 on December 9, 2023. It says the deal was reached on Friday, which was December 8.
Source: Artificial Intelligence Act: deal on comprehensive rules for trustworthy AI, European Parliament
Parliament adopted its position by 499 votes in favor, 28 against and 93 abstentions, before talks with member states. MEPs widened the list of bans to include real-time remote biometric identification in publicly accessible spaces. They wanted generative AI systems to disclose AI-generated content.
Source: MEPs ready to negotiate first-ever rules for safe and transparent AI, European Parliament
The Council, meeting as the Transport, Telecommunications and Energy Council, approved its general approach. The Council document calls this its provisional position on the proposal. It also calls it the starting point for preparing negotiations with the European Parliament.
The Commission published its proposal for a regulation on artificial intelligence. It described the proposal as the first legal framework on AI. The proposal sorted AI uses into four risk levels: unacceptable, high, limited and minimal.
The EU AI Act applies to companies worldwide that sell or use AI in the EU and phases in from 2025 to 2028. A July 2026 amendment moved its high-risk rules to December 2027 and August 2028.
The US has no comprehensive federal AI law. AI is governed by executive orders, agency powers, a few narrow statutes and more than 100 state AI laws enacted in the first half of 2026.