The European Parliament’s IMCO and LIBE committees have now joined the EU Council in proposing significant changes to the EU AI Act’s compliance schedule. That institutional alignment matters. When both chambers of the EU legislative process move in the same direction, the probability of delay shifts from possible to likely.
What the committees are proposing, according to the European Parliament’s March 16 press release, is an extension of the current August 2, 2026 high-risk AI compliance deadline, confirmed by Baker Botts as the operative date for organizations currently planning their compliance programs. According to the MEP committee proposals, the new deadlines would be December 2, 2027, for listed high-risk AI systems and August 2, 2028, for high-risk AI embedded in products covered by EU sectoral safety legislation.
These proposals are not yet law. A full Parliament vote is expected in June 2026, with amendments published in July, according to Burges-Salmon’s analysis. Nothing here cancels your existing compliance obligations until that vote concludes and formal amendments are published.
The more novel development is the nudifier ban. The MEP proposals include a new prohibition on AI systems designed to generate non-consensual explicit images of real people. This category of tool doesn’t appear in the EU Council’s streamlining proposals. It represents MEPs adding scope, not just rescheduling it.
On watermarking, MEPs and the European Commission are not yet aligned. MEPs have proposed November 2, 2026 as the AI-generated content watermarking deadline. The Commission’s proposal is February 2, 2027. That gap will need to close before any amendment becomes final.
The compliance planning implication is clear: teams working toward August 2026 are in a holding pattern, not a reprieve.