Overview
- Investigators are examining whether Grok enabled the spread of illegal, sexualized deepfakes, including images that could constitute child sexual abuse material.
- X limited image generation to paying users and added technical filters, but EU officials say these steps do not sufficiently address systemic risks.
- NGO analyses estimate millions of Grok-generated images in early January, including roughly 23,000 depicting children, figures not yet judicially verified.
- The Commission also widened an ongoing case to scrutinize X’s AI-driven recommendation systems and can demand documents, question staff, and impose fines under the DSA.
- Parallel actions include a UK Ofcom probe, inquiries by Paris prosecutors, and temporary blocks of Grok in Indonesia and Malaysia, with a prior €120 million EU fine against X unrelated to Grok still pending payment.