European Parliament: position on the AI Act
On 26 March 2026 the European Parliament adopted, by 569 votes to 45 with 23 abstentions, its position on simplifying the AI Act: 2 December 2027 and 2 August 2028 as dates for high-risk systems, a ban on 'nudifier' systems, and until 2 November 2026 to mark synthetic content.
Why it matters
Both institutions that have to agree the law now had written positions with the same dates for high-risk systems, so negotiations began from the detail rather than from whether to delay. An editorial assessment.
What the position contains, per the Parliament's press release. Dates: 2 December 2027 for the high-risk systems listed in the regulation (biometrics, critical infrastructure, education, employment, essential services, law enforcement, justice, border management) and 2 August 2028 for systems covered by sectoral safety law. Providers to have until 2 November 2026 to mark AI-created content. A ban on systems that create or manipulate explicit images of an identifiable person without consent; it does not apply where effective safety measures exist. Processing personal data to detect bias only when strictly necessary. Support for small mid-cap enterprises. What the record does not claim. This is a negotiating position with the Council, not law. The adopted text the release links was not opened: the vote count and dates are the press release's.