EU Digital Omnibus reshapes AI Act timelines: high-risk deadline shifts toward December 2027

OneTrust analysis of the in-progress trilogue on the Digital Omnibus on AI reports that both the Council and Parliament are aligned on pushing most high-risk AI obligations from August 2026 to December 2027, with systems embedded in regulated products moving to August 2028. A second milestone requires watermarking of AI-generated content by November 2026. A final agreement is targeted for the 28 April trilogue.

In this domain: Scientific American: cybersecurity experts divided on Mythos severity as Bank of England intensifies AI risk testing · Agentic AI governance gap identified as core EU AI Act compliance challenge · White House in talks with Anthropic over Mythos Preview government access despite Pentagon blacklist