EU AI Act amendments prohibit AI systems that generate nonconsensual intimate imagery and CSAM

Amendments to the European Union AI Act introduce two new prohibited practices taking effect December 2, 2026: a ban on placing on the market, putting into service, or using AI systems that generate or manipulate realistic depictions of an identifiable person's intimate parts or sexually explicit activities without freely given, specific, informed, unambiguous, and explicit consent, and a parallel prohibition on AI systems that generate or manipulate child sexual abuse material as defined under Directive 2011/93/EU. The prohibitions apply to both AI system providers and deployers, though the circumstances triggering liability differ between the two categories, and a narrow without-right defense applies under certain national laws for CSAM-related offenses.

A supply-side prohibition on certain generative models addresses the upstream risk but leaves the downstream distribution problem unsolved: open-weight models, models hosted outside EU jurisdiction, and tools accessed through encrypted messaging channels will continue to generate prohibited imagery regardless of the legal status of their European deployment. The enforcement gap between a prohibited model and a prevented harm is where Guardii operates. The platform's real-time anti-CSAM and anti-sextortion filters detect when AI-generated abuse material enters a child's message stream, identify the coercive or grooming context in which it appears, and surface the conversation to a parent, school safeguarding lead, or law enforcement before the interaction escalates, a targeted intervention that does not depend on whether the generating model was lawfully placed on the market or whether the user uploading it is within reach of EU enforcement.