EU Lawmakers Reach Deal to Ban AI Nudification Tools

On May 7, 2026, the European Parliament, the European Council, and the European Commission reached a provisional agreement in Brussels to streamline the EU AI Act. The deal introduces a strict ban on AI-driven nudification tools and the creation of unauthorized sexually explicit images, a move prompted by public outcry over non-consensual content produced by the Grok chatbot. These protections, championed by lawmakers Michael McNamara and Kim van Sparrentak, will apply to both stand-alone systems and embedded tools.

To reduce administrative burdens on businesses, the agreement significantly delays the enforcement of high-risk AI provisions. The deadline for stand-alone AI systems moves from August 2026 to December 2, 2027, while embedded tools are postponed to August 2028. Additionally, the deal excludes machinery from the Act to remove overlapping safety rules and expands exemptions for small-to-mid-cap companies.

Other key mandates include a December 2, 2026 deadline for mandatory watermarking of AI-generated output and the creation of an EU-level sandbox for developers. The European Commission AI Office will eventually hold enforcement powers and gain access to providers' internal safety practices. The agreement now awaits formal endorsement by EU governments and the European Parliament before becoming law.