EUROPE REWRITES ITS AI LAW: DEADLINES PUSHED BACK 16 MONTHS, COMPANIES GET BREATHING ROOM — BUT DEEPFAKE PORN IS NOW EXPLICITLY BANNED The European Union spent two years building the most comprehensive AI regulatory framework in the world. It has now spent several months quietly taking it apart and putting it back together in a simpler form, with longer deadlines, fewer requirements, and a harder line on the most harmful applications. The Digital Omnibus agreement, finalized May 7 between the EU Council and Parliament, represents the first formal amendments to the AI Act since its adoption in June 2024. The most significant change for businesses is timing. High-risk AI systems covering employment, education, and health insurance now have until December 2027 to comply, a 16-month extension from the original August 2026 deadline. Product-regulated AI systems covering medical devices and similar categories have until August 2028. The changes extend regulatory exemptions previously reserved for small businesses to a broader class of mid-size companies, reduce documentation requirements in targeted areas, and consolidate oversight authority in the EU’s AI Office to cut down on bureaucratic fragmentation across member states. But the agreement added two new prohibitions that were not in the original text. AI systems used to generate or distribute non-consensual intimate imagery are now banned under Article 5, effective December 2026. AI-generated child sexual abuse material faces the same prohibition with the same timeline. The EU is framing this as balance — give industry more runway to comply, while drawing harder lines on the most harmful applications already proliferating across the continent. Keywords: EU AI Act update, Digital Omnibus AI rules, AI regulation Europe 2026, deepfake ban EU