Starmer vs X

A political standoff has emerged between the UK’s Labour government, Elon Musk’s X, and, increasingly, Donald Trump’s White House. What began as a brief and relatively contained trend involving X’s AI tool Grok has escalated into a test case for the UK’s Online Safety regime, a broader battle over free speech, and a flashpoint in transatlantic relations.

At the centre of the controversy was a short-lived trend in which some users prompted Grok to use pre-existing photos of women on the platform to generate images of them wearing bikinis. These images were often directed at female public figures and were clearly intended to provoke a reaction. The practice was swiftly condemned by many users on X and, while unpleasant, it was not as widespread as subsequent political rhetoric suggested.

It is easy, and appropriate, to sympathise with women who found themselves targeted. Being subjected to unwanted sexualisation, even in a relatively mild form such as digitally imposed swimwear, can feel like an invasion of privacy and a loss of agency. The discomfort experienced by the women targeted is perfectly understandable.

As the trend continued, a number of prominent male figures were subjected to the same treatment, including former Venezuelan president Nicolás Maduro, Donald Trump, Keir Starmer, and others. Notably, these instances generated little official concern or acknowledgement. There were also a handful of fringe cases, widely circulated precisely because of their offensive nature, in which users attempted to generate similar images involving minors. These were condemned universally on X, fell squarely within existing criminal prohibitions, and were removed.

Much of the activity around Grok during this period did not involve harassment at all. The vast majority of users prompting the AI to generate bikini images were women editing photos of themselves, often to promote adult content hosted on other platforms. In a paradoxical twist, other users responded by prompting Grok to add clothing to these images rather than remove it. Taken together, the evidence suggests that a small minority of users were exploiting the tool primarily to troll others and provoke outrage.

The Labour government’s response to this emergent trend was swift and uncompromising. Prime Minister Keir Starmer publicly suggested that X could lose its right to self-regulate, stating: “If X cannot control Grok, we will.” The government has also announced plans to bring into force new legislation making it illegal to create non-consensual intimate images. While it is already illegal to share fake intimate images, the proposed law goes further by criminalising the act of creation itself.

Ministers have further indicated that they intend to criminalise the supply of online tools used to generate such material, a move with potentially far-reaching consequences. If implemented broadly, this could affect a number of other AI developers as well as image-editing software such as Photoshop, platforms uninvolved in the current controversy. Technology Secretary Liz Kendall emphasised that the proposed measures were not intended to target Grok specifically, which will be exempt. The policy was announced in December, before the trend on X emerged.

Ofcom, the UK’s communications regulator, has since launched an investigation into whether X breached UK law. Under the Online Safety framework, if violations are found, the regulator has the power to fine the company up to 10% of its worldwide revenue or £18 million, whichever is greater. If the platform fails to comply, Ofcom can seek a court order requiring UK internet service providers to block access to X entirely.

X, for its part, has initially pushed back. In a statement from its official safety account, the platform said it takes action against illegal content, including child sexual abuse material, by removing it, permanently suspending accounts, and cooperating with law enforcement where necessary. The company added that anyone using or prompting Grok to create illegal content would face the same consequences as if they had uploaded illegal material directly.

From X’s perspective, the platform is being held to an impossible standard. No system that allows user-generated content at scale can guarantee the complete absence of abuse. By framing the issue as a failure of moderation rather than an inevitability of scale, the government gains a politically useful narrative. X can be portrayed as reckless, hostile, or uniquely dangerous, even when the underlying behaviours are endemic across the internet as a whole.

This framing has not gone unnoticed in Washington. Sarah B. Rogers, the US Under Secretary of State for Public Diplomacy, stated that “nothing is off the table when it comes to free speech” and suggested that the US would be watching Ofcom’s actions closely. A State Department source told The Telegraph that UK officials could be prevented from entering the US if X is banned. Florida Republican Anna Paulina Luna has gone further, announcing that she is drafting legislation to sanction the UK should such a ban occur. Elon Musk himself has responded in characteristically blunt terms, describing the Labour government as fascist.

What the UK’s Labour government presents as a matter of online safety and harm reduction is viewed by many, particularly in the US, as an attempt to exert state control over speech through regulatory pressure. The fact that the most severe sanctions available include the removal of access to an entire platform only strengthens that perception.

There is also a strategic dimension to the government’s approach. By holding X to standards that are arguably unattainable, it creates a standing justification for reputational and political attacks on a platform that is often hostile to it. This is particularly attractive given that X is the preferred platform for dissidents most critical of the government. The intervention of Technology Secretary Liz Kendall further reinforced the perception that these actions were motivated by political ambition rather than purely by public protection, particularly through urging the regulator not to take “months and months” to reach a conclusion. Regulatory enforcement thus becomes a tool of power.

The internet, however, has never been a polite or sanitised environment. One should expect that every form of behaviour of which human beings are capable will manifest online to some degree. The network connects billions of people across cultures, norms, and moral frameworks, and a non-trivial proportion of those people are hostile, bored, or cruel. There is often a kind of buyer’s remorse among those who build large online platforms or large personal followings. The imagined rewards are visibility, influence, and opportunity. The reality is that all high-profile users, without exception, encounter abuse, trolling, threats, and unwanted sexualisation. None of this makes such behaviour acceptable, but it should be considered predictable.

What is less often acknowledged is that online abuse derives much of its power from constant engagement. For many, the healthiest response is to disengage, to recognise the difference between online discourse and everyday life, and to refuse to allow social media to dictate real-world emotion. This is easier said than done, but it highlights the limits of regulation as a substitute for personal virtue.

None of this is to deny that non-consensual sexual imagery is harmful, or that AI tools introduce genuinely new challenges for legislators. It is, however, reasonable to question whether this particular episode warrants the level of escalation it has received, and whether the response is proportionate to the scale of the problem. There is a significant difference between addressing genuine abuse and attempting to engineer an internet free of unpleasantness. The former is necessary. The latter is impossible.

Following the continued controversy, X has responded by limiting Grok’s image-generation feature to verified profiles. This was a temporary measure while the platform implemented measures that would prevent Grok from generating images of real people in revealing clothing. These actions clearly acknowledge the regulatory concerns and provide a demonstrable response, but whether they will be sufficient to abate the potential consequences of the Ofcom investigation remains to be seen.

The outcome of this online trend will therefore matter far beyond X itself. It will signal how far the UK’s Labour government is willing to go in imposing its moral and political preferences on global platforms, how aggressively regulators are prepared to stretch their mandate, and how much friction the government is willing to tolerate with their closest ally. In that sense, the controversy is not really about bikinis at all. It is about how free the internet will be in the years to come.