EU opens formal probe into X over Grok-generated sexual deepfakes and platform safeguards

Investigation targets deepfakes, risk controls and compliance

The European Union has launched a formal investigation into X, focusing on whether the platform’s AI chatbot Grok enabled the creation and spread of non-consensual sexualized deepfake images. Regulators are examining whether X met obligations under the EU’s Digital Services Act to assess and mitigate systemic risks tied to illegal content and harms to users.

The probe comes amid rising concern in Europe that generative AI features can scale existing online harms faster than platforms can respond, particularly when image manipulation tools are easy to access and content circulates through recommendation systems.

How Grok’s image tools became the center of controversy

Reports described Grok generating or enabling sexually explicit image manipulations, including alleged material involving women and possibly minors. European officials said the investigation will look at whether the platform’s safeguards, reporting tools and enforcement mechanisms were sufficient given the risks associated with synthetic media.

Part of the concern is structural: when a system makes outputs visible or shareable by default, harmful content can spread rapidly before moderators can identify and remove it. The EU’s inquiry is expected to examine not only the tool itself but also the broader distribution dynamics on X.

Recommender systems and AI-driven curation in the spotlight

Regulators are also scrutinizing X’s recommendation systems, especially as the company moves to integrate Grok more deeply into the platform. Under the DSA, very large online platforms can face significant obligations related to transparency, risk assessments and mitigation measures when algorithmic amplification contributes to harmful outcomes.

The case may set an important precedent for how Europe enforces platform accountability when generative AI is embedded into social products. Depending on findings, outcomes can include mandated product changes and potentially large penalties.

What platforms may be pushed to do next

• Hardened guardrails and stricter limits on editing real-person images
• Faster reporting and takedown workflows for non-consensual content
• Greater transparency about how AI features and recommendations function
• Investment in provenance, watermarking and authenticity tooling

X has said it maintains policies against child sexual exploitation and non-consensual nudity and that it is committed to user safety, but European officials indicated they will judge compliance based on concrete, audited controls and outcomes.