EU’s Digital Omnibus on AI takes effect, delaying high risk rules to 2027

12:01 AM CET · July 30, 2026

The European Union’s Digital Omnibus on AI entered into force this week, three days after publication in the Official Journal, formally delaying some of the AI Act’s most consequential compliance deadlines. Rules for stand alone high risk AI systems listed in the law’s Annex III, including systems used in employment and education decisions, now apply from December 2, 2027 rather than August 2, 2026, while high risk AI systems embedded in regulated physical products get until August 2, 2028. The Commission said the delay was prompted by practical implementation problems, including slow designation of national competent authorities and the absence of finalized harmonized technical standards that companies would need to demonstrate compliance. Most of the AI Act’s transparency obligations, including rules requiring disclosure of AI generated content, remain on track to apply from August 2, 2026, and the omnibus also extends support measures originally aimed at small businesses to small mid cap companies and expands access to regulatory sandboxes, including a new EU level sandbox. The move follows months of industry lobbying arguing the original timeline was unworkable, and it lands the same week several companies, including Meta, formally signed the bloc’s separate code of practice on labeling AI generated content, suggesting firms are moving to meet the transparency rules even as the heavier compliance regime is pushed back.