Signal Why the EU rewrote its landmark AI law / AI Act reloaded

Summary

On 7 May 2026 EU negotiators reached a provisional political agreement on the 'Digital Omnibus on AI,' a package of targeted amendments to the EU AI Act, which the Council formally confirmed with final green light in late June 2026. The co-legislators added a new prohibition to Article 5 targeting AI systems used to generate non-consensual sexual or intimate images of real identifiable persons ('nudification' or deepfake-nude tools) and AI-generated child sexual abuse material (CSAM). The prohibition covers systems where such generation is the intended purpose as well as systems where it is a reasonably foreseeable, reproducible outcome absent adequate technical safeguards, and it also bars deployers from using any system to generate such content regardless of the system's design. Compliance with this prohibition is envisaged by 2 December 2026, and a safe harbour exists for providers with effective preventive safeguards. As part of the same package, high-risk AI system obligations were postponed by over a year: stand-alone Annex III systems must now comply by 2 December 2027 and AI embedded in regulated products under Annex I by 2 August 2028, replacing the Commission's originally proposed conditional trigger mechanism with fixed dates.