EU Commission opens Official Enquiry into X over Grok AI deepfakes

The Grok backlash was never really about one viral prompt. It has always been about what happens when a tool that can “edit” reality is connected to a platform meant for instant distribution, then shipped faster than the safeguards can catch up.

The European Commission made this tension official on January 26. The EU opened a Digital Services Act investigation into X over the sexually explicit deepfakes connected to Grok, including images the Commission says might involve children. The investigation is framed as a question of platform risk, not just individual moderation failures.

The filing also includes a second track that might end up being more consequential than the images. X has signaled it wants Grok to play a role in curating what users see on the timeline. When an AI starts influencing the feed itself, the consequences of mistakes do not stay local.

The new EU investigation into Grok on X

The Commission’s core question is straightforward. Did X properly mitigate the risks of Grok’s image tool when it was deployed inside the EU, particularly the risk of non-consensual sexualized content spreading widely once it hits the feed?

The EU’s stance here differs from the typical “fix it after it breaks” cycle. The Digital Services Act is built on the assumption that retroactive patches are not a defense. The Commission wants to know what risk assessment was done before rollout, if the safeguards were adequate, and if the system was designed to prevent abuse from being easy in the first place.

Why the timeline angle is the real pressure point

In addition to the deepfake probe, the Commission expanded its scrutiny of X's recommender systems following the platform's move to integrate Grok for content curation. This is the part of the code with the widest blast radius.

If a system cannot be trusted to generate images without spawning illegal content, why should it be trusted to decide what news you see next? The EU’s logic is simple: If Grok is too unstable to draw a picture, it is certainly too unstable to curate the “news” on your timeline.

What X has done differently since the backlash

Since the abuse surfaced, X says it has tightened controls, including restricting image editing in jurisdictions where it is illegal. The platform also points to its policies regarding child exploitation and non-consensual nudity.

But those changes are exactly the point of contention. X’s defense is that it fixed the problem fast. The EU’s counter-argument is that under the DSA, you don't get credit for putting out a fire you started. The issue isn't the patch; the issue is the launch, and whether X did the necessary risk work before shipping a high-impact feature into a viral network.

The investigation is part of a wider enforcement push

The timing of the EU actions follows a broader trend. The Grok controversy has prompted international responses, which include reports of restrictions or blocks in countries like Malaysia and Indonesia.

While Europe’s move lands in a broader period of pressure for X. In December 2025, the Commission fined the platform €120 million under the DSA for transparency breaches. This context matters because it tells the reader this isn't a one-off scolding. It is a regulator that has already shown it will use the DSA’s sharp edges.

The evidence hold that signals escalation

One detail to watch is the preservation of evidence. Earlier this month, the Commission ordered X to retain internal documents and data related to Grok through the end of 2026.

Think of it as the EU freezing the internal record. Europe is preserving the paper trail so investigators can answer the blunt question at the heart of negligence cases: what did X know about Grok’s risks, when did it know it, and what did it ship anyway?

The headline controversy began with what Grok could generate. The EU is now forcing a broader question about what the platform can safely distribute. In that framing, the images are the spark, but the timeline is the accelerant.