The UK, Canada and Australia are reportedly in discussions about a possible ban on X due to the explosion of explicit deepfakes that users are doing through Grok. In the text, it is presented as a coordinated idea, and as ‘evidence’, posts on social networks and citations from other media are mostly recounted, so this should be read as a report on alleged conversations, not as a confirmed decision of the three governments. In other words, for now it is a signal of political pressure, not a ready-made measure.
What is more concrete for the European context is that the European Commission is moreć reached for a tool from DSA. Reuters reported on January 8 that the commission had ordered Xu to retain all internal documents related to Grok until the end of 2026 while it reviews compliance, after the commission publicly condemned the sexualized AI images. This is not a ban, but it is a clear sign that the case is being treated seriously and that the EU is gathering material for possible further steps.
The UK, which is not in the EU, but is still a European market that X cannot ignore, is going in the same direction. On January 12, Ofcom opened a formal investigation against X under the Online Safety Act and is seeking explanations and evidence about how the platform prevents illegal content and protects users. It is, in contrast to the ‘banned story’, an official regulatory procedure with very clear powers and deadlines. It is also interesting to note the public spat between the British Prime Minister and Elon Musk, who accused the British government of being fascist because they wanted to abolish freedom of speech.
And then comes a detail that sounded incredible to many, but was announced through multiple sources: US Defense Secretary Pete Hegseth said that Grok will be integrated into the Pentagon’s networks later this month, both on classified and some unclassified systems. This confirms that the American government does not share the concerns of its European colleagues about privacy and the possibility of Grok abuse, and takes us back about a year, when Musk was part of the American executive branch. The Guardian and Defense One write that this is part of a broader ‘AI acceleration’ story in which the Pentagon wants to accelerate the use of more large-scale AI models within its systems, with an emphasis on data availability. In practice, this means that tools that are under scrutiny due to security and abuse will find their way into the most sensitive environments, which will keep this topic in focus for a long time.