About

EU Parliament Votes to Delay High-Risk AI Act Deadlines by 16 Months

Published
Score
8

Why it matters

The European Parliament has voted to extend the EU AI Act's main compliance deadline for high-risk AI systems from August 2, 2026, to December 2, 2027—a 16-month reprieve. The amendment, part of the broader "AI Act Omnibus" simplification package, creates staggered deadlines by AI category: high-risk standalone systems must comply by December 2, 2027, while AI integrated into regulated products like medical devices and machinery must comply by August 2, 2028. The extension grants companies additional time to develop technical standards and regulatory guidance before enforcement begins. The legislation also introduces an immediate prohibition on AI systems generating non-consensual intimate imagery and child sexual abuse material, effective December 2, 2026, covering both tool builders and deployers.

The European Parliament, Council of the European Union, and European Commission reached provisional agreement on these amendments on May 7, 2026. The delay resolves a longstanding "double-compliance" issue where AI-enabled machinery previously had to satisfy both the AI Act and existing product safety regulations simultaneously. A separate transparency requirement for watermarking AI-generated content has been shortened to a three-month delay, with compliance required by December 2, 2026.

Companies deploying AI in high-risk sectors—employment screening, education, law enforcement, critical infrastructure, and biometrics—should treat the December 2027 deadline as their actual compliance target, not the original August 2026 date. The extended timeline provides a critical window to audit systems and governance frameworks, but the prohibition on non-consensual intimate imagery takes effect this December. Organizations should prioritize understanding how the revised deadlines apply to their specific AI deployments and begin alignment with emerging technical standards now, rather than waiting for final regulatory guidance.

Sources

mail Subscribe to Artificial Intelligence email updates

Primary sources. No fluff. Straight to your inbox.

Also on LawSnap