Overview
- Parliament’s civil liberties and internal market committees approved a position to prohibit AI tools that generate sexually explicit or intimate images of identifiable people without consent, with an exception for systems that effectively prevent such outputs.
- The proposed prohibition forms part of targeted amendments to the AI Act and follows backing by EU member states last week.
- If the plenary endorses the mandate on March 26, lawmakers will open talks with the Council to agree a final text before the changes can become law.
- MEPs supported fixed later application dates for some obligations—December 2, 2027 for listed high‑risk uses and August 2, 2028 for AI covered by sectoral product‑safety laws—and a shorter extension to November 2, 2026 for watermarking of AI‑generated content.
- The push follows January outrage over Grok‑generated deepfakes, with the European Commission investigating Grok under online‑content rules and X saying it introduced countermeasures and has zero tolerance for sexualised deepfakes.