EU Digital Omnibus Proposes Delaying AI Act High-Risk Compliance by Up to 16 Months

The European Commission's Digital Omnibus proposal, presented on November 19, 2025, would significantly delay several key compliance deadlines in the EU AI Act — the landmark regulation that was supposed to set a global standard for AI governance. The proposed delays affect both transparency obligations and high-risk AI system requirements, and are now working through European Parliament and Council review.

What Would Be Delayed

AI transparency obligations — requiring disclosure when content is generated by AI for synthetic audio, images, video, and text — were scheduled to take effect for systems placed on the market before August 2, 2026. The Omnibus would delay these obligations by six months to February 2, 2027. For high-risk AI systems, the delays are more substantial: compliance deadlines would be pushed back by up to 16 additional months beyond the original August 2026 timeline.

The high-risk delays follow a two-stage structure. Entry into force for Annex III high-risk systems (covering areas like employment, credit, education, and law enforcement) would begin six months after the Commission confirms that adequate compliance support is available. Entry into force for Annex I high-risk systems would follow 12 months after that confirmation. The net effect is that high-risk AI compliance, originally expected to be enforceable in 2026, may not take effect until late 2027 or 2028 depending on when the Commission certifies its support infrastructure.

Why the Delays

The AI Act was designed to be the world's first comprehensive AI regulatory framework, and its ambition exceeds the readiness of the compliance infrastructure needed to implement it. Regulated organizations need conformity assessment procedures, standards bodies need to publish harmonized standards, and national authorities need to establish the expertise to oversee compliance. None of these prerequisites are fully in place on the original timeline.

Industry lobbying has been intense. AI companies and industry associations argued that the original deadlines were unrealistic and would put European AI development at a competitive disadvantage relative to the US and China, where regulatory constraints are lighter. The Omnibus delays are a practical acknowledgment that the gap between regulatory ambition and implementation readiness is wider than anticipated.

Implications

For AI companies selling into the EU market, the delays provide additional runway to develop compliance programs. However, the delays also create uncertainty: the final timeline depends on when the Commission certifies its compliance support, a date that is not yet fixed. Companies must decide whether to plan for the original deadlines, the proposed delayed deadlines, or some intermediate scenario.

For AI governance advocates, the delays signal that even the most committed regulatory jurisdiction is finding it difficult to translate AI regulatory principles into enforceable rules on the originally planned schedule. The EU AI Act remains the most comprehensive AI regulation anywhere in the world, but its implementation timeline is now subject to the same political and practical pressures that have delayed other ambitious technology regulations.