Brussels: The European Union has given its final green light to a slimmed-down version of the AI Act, pushing back several key deadlines and trimming obligations that industry warned were arriving before the tools to meet them existed.
The changes come through the Digital Omnibus, a package of targeted amendments that the Council and Parliament agreed provisionally on 7 May 2026 and that the Council waved through on 29 June. Formal adoption is expected before 2 August, the date on which the next tranche of the original law would otherwise have started to bite.
The headline shift is time. Stand-alone high-risk AI systems, those deployed in sensitive areas such as hiring, credit and policing, now have until 2 December 2027 to comply, while high-risk systems built into products as safety components get until 2 August 2028. Negotiators justified the delay on practical grounds, noting that many of the technical standards meant to guide compliance are themselves behind schedule.
The omnibus also tidies up where the rulebook overlaps with others. Machinery products will in most cases only have to satisfy sectoral safety rules rather than both those and the AI Act, and systems that merely assist users or optimise performance will not automatically be branded high-risk unless their failure threatens health or safety.
Simplification did not mean softening across the board. Lawmakers added a fresh prohibition on AI systems that generate child sexual abuse material or non-consensual intimate imagery, with compliance required by 2 December 2026, and set the same date for delayed rules on watermarking AI-generated content. The package also reinstates mandatory registration of certain high-risk systems in a central EU database.
Reaction has split along familiar lines. Industry groups welcomed the breathing room, arguing that firms cannot comply with standards that do not yet exist, while digital rights campaigners fear that every postponement chips away at protections the law was written to guarantee. The Commission frames the exercise as smarter regulation rather than retreat.
The full set of changes sits on the Commission’s AI regulatory framework pages. The harder question the amendments cannot answer is whether Europe can police powerful AI systems while the timetable for doing so keeps sliding.
