EU delays high-risk AI Act to 2027

Social posts saying the European Union has pushed “high-risk AI Act” enforcement to 2027 overstate what the bloc has actually published. The European Commission’s implementation timeline says the “majority of rules” start on August 2, 2026, including rules for high-risk AI systems in Annex III and enforcement at national and EU level. A later date — August 2, 2027 — appears on the Commission-backed timeline for high-risk AI embedded in regulated products, with a note that the Commission has proposed linking that step to the availability of support tools and harmonised standards. ### Which part of the AI Act is actually on a 2027 clock? The Commission-backed AI Act Service Desk says August 2, 2027 is the date for “high-risk AI embedded in regulated products.” The same page separately says August 2, 2026 is when “the majority of rules” come into force, including Annex III high-risk systems, transparency rules and enforcement. (ai-act-service-desk.ec.europa.eu) Regulation (EU) 2024/1689 is already in force. The law entered into force on August 1, 2024, according to the European Commission, and the official text was published in the EU’s Official Journal in July 2024. ### Does this mean the EU formally delayed all high-risk AI obligations? The answer from the official EU timeline is no. The public timeline does not say the bloc has postponed all high-risk AI obligations to 2027; it splits the timetable between Annex III high-risk systems in 2026 and high-risk AI embedded in regulated products in 2027. (ai-act-service-desk.ec.europa.eu) (commission.europa.eu) The only explicit reference to a possible change on that page is narrower. The note says that, “in the context of the Digital Omnibus package,” the Commission has proposed linking the application of rules for high-risk AI systems to the availability of support tools, including harmonised standards. The page does not describe that as a blanket delay for every high-risk category. (ai-act-service-desk.ec.europa.eu) ### What are Annex III systems, and why does that distinction matter? The Commission’s AI Act FAQ separates high-risk AI questions from the broader implementation timetable and specifically points readers to Annex III use cases when asking whether a system is high-risk. That matters because the implementation page assigns Annex III systems to the August 2, 2026 phase, not the 2027 phase. (ai-act-service-desk.ec.europa.eu) The law itself creates multiple buckets of obligations that start at different times. The Commission’s published rollout already included staggered dates: prohibitions from February 2, 2025, general-purpose AI obligations from August 2, 2025, most AI Act rules from August 2, 2026, and the regulated-product category from August 2, 2027. (digital-strategy.ec.europa.eu) ### Why are people saying “delay” now? A December 4, 2025 Commission update helps explain the confusion. The AI Office said it was preparing a long list of 2026 guidance documents on high-risk classification, transparency, serious-incident reporting, provider and deployer obligations, substantial modification, post-market monitoring and the law’s interaction with other EU rules. (ai-act-service-desk.ec.europa.eu) That same update said the work was being advanced “in the context of the implementation of the AI Act and the proposed Digital Omnibus.” Read together with the implementation timeline, the Commission is signaling that support materials and simplification work are still being built out, but the official public timetable still keeps Annex III high-risk obligations on the 2026 date. (digital-strategy.ec.europa.eu) ### So what should companies use as the operative dates right now? August 2, 2026 is still the operative date on official EU pages for Annex III high-risk systems and for the start of most AI Act enforcement. August 2, 2027 is the date shown for high-risk AI embedded in regulated products, alongside the Commission’s proposal to tie that step to support tools and standards. (digital-strategy.ec.europa.eu) In 2026, companies looking for the next concrete step should watch the European Commission’s AI Office guidance pages and the AI Act Service Desk. The Commission has said guidance on high-risk classification and obligations is due in the course of 2026, and the Service Desk timeline is where the bloc is publishing the staged rollout dates. (digital-strategy.ec.europa.eu) (ai-act-service-desk.ec.europa.eu)