Digital Omnibus on AI (Regulation (EU) 2026/1744)

The Commission introduced the AI-specific strand of its Digital Omnibus package on 19 November 2025. The Council and Parliament reached a provisional agreement in May 2026, the Council gave final approval on 29 June, and the act was signed and dated 8 July. Its enactment changed the relevant reference point: proposal-stage summaries should now be checked against Regulation 2026/1744 and the consolidated AI Act.

Sources: s1, s2, s3, s4, s5

The amendment moves the application of the specified Chapter III high-risk provisions to 2 December 2027 for systems classified under Article 6(2) and Annex III, and to 2 August 2028 for systems classified under Article 6(1) and Annex I. This is not a postponement of every AI Act obligation. The law also changes sandbox arrangements, some administrative requirements, and AI Office supervision. Newly added prohibitions concerning non-consensual sexual or intimate content and child sexual abuse material have a separate application date of 2 December 2026.

Sources: s1, s2, s3, s5

Consider a team comparing an employment-related AI system with a safety component in a regulated product. Both may raise high-risk questions, but they do not necessarily share an application date or classification route. A useful reading separates three questions: which category applies, which amended provision imposes the obligation, and whether a transition rule changes its timing. A blanket claim that the AI Act has been delayed until 2028 would erase those distinctions. This comparison illustrates how to read the timetable; it does not classify a particular product.

Sources: s1, s3

This summary is not legal advice and cannot determine whether a specific system is prohibited, high-risk, or subject to a particular transition date. The regulation must be read with the consolidated AI Act, sectoral law, implementing measures, and current guidance. Proposal commentary may be obsolete where negotiations changed the text.

Sources: s1, s3, s4, s5