EU agrees ban on AI systems generating sexualised deepfakes
The European Union has agreed to ban artificial intelligence systems that generate sexualised deepfakes, in a move that will be added to changes to the bloc's 2024 AI rules.
The decision follows concern over non-consensual nude images produced by AI tools.
Irish Independent MEP Michael McNamara said the EU had drawn a red line and that the legislation explicitly bans nudifier applications.
EU negotiators from the European Parliament and European governments also agreed to delay the rollout of some high-risk AI rules.
Those rules had been due to apply from August 2026 for stand-alone systems and a year later for AI tools embedded in other products.
Under the revised timetable, the start dates will move to December 2027 and August 2028 respectively.
The EU executive had proposed the amendments last year, saying they were intended to help businesses and avoid stifling innovation.
The change matters because it adds a specific prohibition to the EU's wider AI framework while also slowing parts of the regime for systems judged potentially dangerous to safety, health or fundamental rights.
The AI regulation office, which the EU says will have access to providers' internal safety and security practices, is due to gain enforcement powers in August 2026.
