EU moves to ban malicious AI images

The European Parliament’s Civil Liberties (LIBE) and Internal Market (IMCO) committees approved an amendment on March 18, 2026 explicitly outlawing AI systems that generate realistic sexually explicit or intimate images of identifiable people without consent. (bloomberg.com) The Council of the EU adopted a mandate on March 13, 2026 that inserts a prohibition on “generation of non‑consensual sexual and intimate content or child sexual abuse material” into the AI Act and set new application dates of 2 December 2027 for stand‑alone high‑risk AI and 2 August 2028 for high‑risk AI embedded in products. (consilium.europa.eu) Parliamentary committees also voted to postpone parts of the high‑risk AI timetable and proposed a shorter delay — moving some implementation deadlines to 2 November 2026 instead of the Commission’s suggested 2 February 2027 — creating a clear timing dispute ahead of trilogue talks. (rte.ie) The legislative push followed the Grok controversy earlier this year, during which watchdogs and NGOs estimated millions of sexualized images were generated and shared, and Ireland’s Data Protection Commission opened a formal inquiry on February 17, 2026. (counterhate.com) The UK Information Commissioner’s Office announced investigations into X and xAI on February 3, 2026, and California’s Attorney General launched a probe in mid‑January 2026, signalling parallel regulatory scrutiny outside the EU. (ico.org.uk) If the ban is adopted, providers of advanced generative models will be required to demonstrate technical measures and restrictions on image generation, and the Council text also reinstates obligations to register high‑risk AI systems in the EU database. (bloomberg.com)