EU opens formal investigation into X over Grok deepfake sexual images
European regulators have opened a new Digital Services Act case targeting X after its Grok chatbot was linked to nonconsensual sexualized deepfakes. The inquiry also widens scrutiny of how X’s recommendation systems may amplify harmful synthetic content.
What the EU is investigating
European Union regulators on Monday, January 26, 2026 opened a formal investigation into X, focusing on whether the platform’s Grok AI tool has complied with the bloc’s Digital Services Act (DSA) after concerns that Grok could be used to generate and spread nonconsensual sexual deepfake images. Officials say the probe will examine whether X properly assessed and reduced systemic risks tied to illegal and harmful content, including manipulated sexually explicit images that may involve children.
The case follows a wave of backlash after users demonstrated how the chatbot’s image capabilities could be used to “digitally undress” people or place them into sexualized scenarios without consent. Regulators are looking at the safeguards X had in place, how quickly the company responded, and whether its mitigation measures were adequate once the risk became apparent.
Why recommender systems are part of the case
The EU also said it is extending a separate, ongoing DSA investigation into X’s recommendation systems, reflecting concern that algorithmic distribution can accelerate the reach of harmful synthetic media. Officials have argued that even if a platform removes some content, the design of engagement-driven ranking can still enable rapid reuploads, resharing, and amplification.
DSA investigations can end with commitments to change product design, moderation workflows, and transparency practices, or escalate to large penalties if regulators conclude the platform failed to meet its obligations. In this case, the EU is explicitly framing nonconsensual sexual deepfakes as a serious safety and rights issue, with heightened attention to protection of women and children.
X’s response and what comes next
X has said it has zero tolerance for child sexual exploitation and nonconsensual nudity, and the company has pointed to changes that restrict certain image-generation or editing behaviors in places where such depictions are illegal. The EU’s investigation, however, will focus on whether the platform’s systems and enforcement were strong enough to prevent harms from materializing and spreading in the first place.
The outcome will be closely watched by other major platforms and AI developers because it signals how aggressively regulators intend to police generative tools integrated into large social networks. It may also shape future expectations for provenance, watermarking, age-safety controls, and rapid-response takedown pipelines for synthetic abuse.