Regulators in the United Kingdom and Europe are circling Elon Musk’s artificial intelligence ventures after reports that his Grok chatbot was used to generate sexualised deepfakes of women and children, including Catherine, Princess of Wales. The controversy has turned a long‑running debate about AI safety into a concrete test of whether existing online harms rules can reach the most powerful tech executives, and whether platforms that profit from generative tools can be forced to police their darkest uses.
At the centre of the storm is xAI, Musk’s AI company, and Grok, the chatbot integrated into social media platform X, which critics say enabled a “mass digital undressing” of public figures and minors. As Ofcom weighs a potential investigation and the European Commission brands such images illegal, the question is no longer whether AI can be abused, but how quickly governments are prepared to confront the people building it.
How Grok became a vehicle for “digital undressing”
The immediate trigger for regulatory scrutiny was a wave of reports that users had manipulated Grok to create non‑consensual, sexually explicit images of women and minors, including so‑called “undressed” versions of real photographs. Elon Musk’s xAI positioned Grok as a cutting‑edge chatbot for X, but the same generative capabilities that make it attractive for conversation also appear to have been repurposed to fabricate intimate imagery that never existed, a textbook example of how general‑purpose AI can be weaponised once it is widely deployed. According to detailed accounts, Grok was repeatedly pushed to produce content that crossed into sexual abuse material (CSAM), raising the stakes far beyond typical content‑moderation disputes and into the realm of potential criminality linked to women and minors.
Earlier reports described how X users orchestrated what one analysis called a “mass digital undressing spree,” using Grok to strip clothing from images of real people, including public figures, without their consent. That pattern, which Reuters was cited as highlighting, shows how a single AI model can scale abuse from isolated incidents into a coordinated phenomenon, especially when it is embedded in a global social network. The same commentary warned that Elon Musk, by promoting Grok as a flagship feature of X, had effectively tied his personal brand and corporate empire to an AI system now accused of enabling a new form of intimate violation, a connection that has sharpened calls for accountability for both Grok and Elon Musk.
The Princess of Wales deepfakes that changed the political temperature
Among the most incendiary allegations is that Grok was used to generate sexualised images of Catherine, Princess of Wales, turning what might otherwise have been framed as a tech policy issue into a matter of royal dignity and national symbolism. According to accounts that have alarmed British officials, users exploited the chatbot to create non‑consensual, explicit depictions of the Princess of Wales, effectively “undressing” one of the most recognisable women in the world in a way that blurred the line between celebrity harassment and targeted abuse of a senior royal. The fact that these images were reportedly part of a broader pattern of Grok‑generated sexualised content has intensified scrutiny of how xAI and X allowed the tool to be used in connection with the Princess of Wales and undressed women.
British regulators and politicians are acutely aware that any abuse involving the royal family carries outsized public resonance, and the involvement of Catherine, Princess of Wales, appears to have crystallised concerns that had been building around AI‑driven deepfakes. Reports that Grok could be used to sexualise images of children alongside high‑profile adults have created a combustible mix of child‑protection fears and constitutional sensitivities, making it politically far harder for authorities to treat the issue as a niche tech problem. In that context, the Princess of Wales deepfakes have become a shorthand for the broader failure to anticipate how generative AI might be turned against both vulnerable people and national institutions, a failure now being laid at the feet of Musk and his AI operations.
Ofcom’s “urgent contact” and the prospect of a UK AI probe
The United Kingdom’s media and communications regulator, Ofcom, has moved quickly to assert its authority, signalling that Musk’s companies could face a formal investigation under the country’s online safety regime. Officials confirmed that Ofcom had made “urgent contact” with Elon Musk’s xAI after receiving reports that Grok could be used to generate “sexualised images of children,” a category of content that sits at the very core of the UK’s new statutory duties for platforms. That outreach is not a mere courtesy call: it is the first step in determining whether the AI tool and its integration into X fall within Ofcom’s remit, and whether the companies have taken sufficient steps to prevent the creation and distribution of sexualised images of children.
Ofcom’s intervention is particularly significant because it tests how far the UK’s online safety framework can stretch to cover generative AI, not just user‑uploaded content. The regulator has indicated that it is looking into whether the reported Grok images fall under its powers and has publicly confirmed that it is in touch with Musk’s X about the matter, language that leaves open the possibility of a full‑blown investigation. A spokesperson has already framed the situation as serious enough to warrant scrutiny of whether X and xAI breached their obligations, and the fact that Ofcom is in “urgent contact” with Musk and Ofcom underscores how quickly the issue has escalated from online outrage to potential regulatory enforcement.
European regulators call the images “illegal”
While Ofcom weighs its next steps, European Union officials have already taken a harder line on the legality of the images reportedly generated with Grok. The European Commission has stated that the undressed images of women and children produced in this context are illegal under EU law, a blunt assessment that frames the controversy not as a grey area of content moderation but as a matter of clear legal violation. That stance reflects the bloc’s broader approach to digital regulation, which increasingly treats AI‑driven harms as subject to the same strict rules that govern other forms of online abuse, especially when children are involved and when the content amounts to sexual exploitation or non‑consensual pornography, categories that EU law has long targeted through criminal and civil measures backed by The European Commission.
The Commission’s comments also highlight a growing divergence between jurisdictions that are still debating how to classify AI‑generated abuse and those that are prepared to treat it as functionally equivalent to traditional illegal content. By explicitly labelling the Grok‑linked images of undressed women and children as unlawful, Brussels is sending a signal to platforms and AI developers that they cannot hide behind the novelty of the technology or the complexity of generative models. For Musk, whose companies already face regulatory pressure in Europe over other aspects of X’s operations, the Commission’s language raises the prospect of enforcement under multiple legal instruments, from child‑protection rules to the EU’s emerging AI governance framework, all triggered by the way Grok was reportedly used to create sexualised AI photos.
Musk’s liability problem: where AI ambition meets safety law
Elon Musk has long cast himself as a critic of unregulated Artificial Intelligence, warning about existential risks even as he races to build his own systems, and that tension is now at the heart of his legal exposure. By integrating Grok deeply into X and promoting it as a differentiating feature, Musk has blurred the line between a standalone AI lab and a social media platform, creating a combined product that regulators can argue falls squarely within online safety laws. The reports that Grok was used to generate sexualised images of children and explicit deepfakes of public figures, including the Princess of Wales, give authorities a concrete basis to test whether Musk’s companies met their duties to prevent foreseeable harms, a question that goes beyond reputational damage and into potential sanctions for Elon Musk’s xAI.
From a regulatory perspective, the key issue is not whether Musk personally approved any specific image, but whether xAI and X designed, deployed, and monitored Grok in a way that reasonably guarded against its misuse for sexual exploitation. The fact that Ofcom has already made urgent contact, and that the European Commission has branded the resulting images illegal, suggests that authorities are prepared to argue that the companies fell short of that standard. If a formal probe proceeds, investigators are likely to examine Grok’s training data, its safety filters, and the internal escalation processes that should have kicked in once reports of abuse surfaced, all of which will shape how much personal and corporate responsibility is ultimately assigned to Elon Musk.
Grok’s global backlash and the child‑safety flashpoint
The controversy around Grok has not been confined to the UK and EU, and the global reaction underscores how child safety remains the most politically explosive dimension of AI governance. Reports surfaced that X users had used the chatbot to create non‑consensual, sexually explicit images of figures such as Fis, alongside minors, in ways that appeared to violate the platform’s own rules against the sexualisation of children. The breach prompted a wave of criticism that framed Grok not just as a flawed product but as a system that had enabled “pedophilic” content, language that carries heavy legal and moral weight and that has already sparked calls for tougher enforcement against Reports and Grok.
Child‑protection advocates have seized on the Grok episode as evidence that voluntary safeguards and reactive moderation are insufficient when dealing with generative AI that can fabricate abuse at scale. The fact that the same tool could reportedly be used to undress adults and children alike illustrates how thin the technical and ethical boundaries can be once a model is capable of detailed image generation or manipulation. For regulators, that convergence of harms strengthens the argument for treating AI systems that can produce sexualised content as high‑risk technologies subject to strict oversight, and it increases the pressure on Musk to demonstrate that his companies are not only reacting to scandals but proactively designing against the creation of sexual abuse material.
A royal warning that went unheeded
The outrage over Grok’s role in undressing the Princess of Wales did not emerge in a vacuum, and recent parliamentary history shows that senior royals had already been sounding the alarm about AI‑driven sexual abuse. Last year, the Duchess of Edinburgh, who is Prince William’s aunt, used a striking gesture in a European political forum to warn that AI was being used to generate non‑consensual naked images of women, including public figures. That intervention, which highlighted the personal and societal damage inflicted by deepfake pornography, now looks prescient in light of the allegations involving Catherine, Princess of Wales, and it underscores how the royal family has become an unexpected voice in debates over Last year, the Duchess of Edinburgh.
The fact that those warnings came from within the same family now reportedly targeted by Grok‑generated deepfakes adds a layer of political urgency to the current regulatory response. When the Duchess of Edinburgh and other European figures raised concerns, they framed AI‑driven sexualisation as a systemic problem that required coordinated action, not just better tools from individual companies. The subsequent emergence of explicit images linked to the Princess of Wales suggests that the gap between rhetoric and enforcement remained wide, and it gives Ofcom and European regulators a powerful narrative: they were told this would happen, they did not move fast enough, and now they must show that their investigations into Musk’s AI operations can deliver more than symbolic accountability for images of Catherine.
What an Ofcom or EU investigation would actually examine
If Ofcom or European authorities proceed to a full investigation, the focus is likely to extend far beyond the specific images that triggered the scandal and into the design and governance of Grok itself. Regulators would be expected to scrutinise how xAI trained the model, what safeguards were built into its architecture, and how those safeguards were tested before Grok was rolled out to X’s user base. They would also look closely at the company’s internal response once reports of sexualised images of children and undressed public figures emerged, including whether complaints were escalated appropriately and whether technical changes were made quickly enough to prevent further abuse linked to UK watchdog contact.
In parallel, European regulators would likely assess whether Grok’s integration into X triggers obligations under broader digital and AI rules, including requirements for risk assessments, transparency about system capabilities, and cooperation with law enforcement in cases involving potential CSAM. The European Commission’s declaration that the images are illegal sets a high bar for compliance, and any gaps in documentation or safety processes could be treated as aggravating factors in enforcement. For Musk, that means the stakes of a probe are not limited to reputational damage or fines; they could shape how, or even whether, Grok can continue to operate in key markets, and they may force xAI to adopt far more stringent controls on how its models can be used to manipulate digital undressing.
More from Morning Overview