go to the original language article
BRUSSELS – Grok is under fire because it creates thousands of AI-generated images per hour of people who have been undressed without their consent. This also involves sexually explicit images of children that are being distributed via X. Earlier, the Commission called this “disgusting” and “illegal” and said it would look into the complaints.
The fact that the Commission is asking X to preserve the documents does not yet mean that an official investigation has been launched into compliance with EU digital legislation. “This is a message to a platform: keep your internal documents, do not throw them away, because we have doubts about your compliance. And we must be able to gain access to them if we explicitly request it,” said the spokesperson.
France and Malaysia have already launched an investigation into Grok.
Elon Musk’s chatbot Grok generates about 6,700 images per hour of people who, without their consent, have been undressed by AI. This emerges from research by a deepfake researcher. The five websites that are then most used for such requests generate an average of 79 images per hour.
Since the end of December, users of Musk’s platform X have been able to use Grok to edit an image. This has already caused an uproar over the past week. Grok, for example, replaces people’s clothes with a bikini on request.
The scale of deepfakes on X is “unprecedented,” said a lawyer specializing in online sex crimes to news agency Bloomberg.
(8 January 2026)