Elon Musk’s artificial intelligence chatbot Grok is at the center of a storm over sexually explicit image manipulation and a new paywall that critics say solves nothing. After viral “undressing” deepfakes triggered political outrage, X responded by locking Grok’s image tools behind a subscription, a move that has been condemned as both cosmetic and offensive to the people harmed.
The backlash now stretches from technology researchers to senior European politicians, who argue that charging for access to risky AI features looks less like safety and more like monetizing abuse. The controversy is rapidly becoming a test case for how platforms handle generative image tools that can be weaponized against public figures and ordinary users alike.
From viral deepfakes to a paywalled “fix”
The immediate trigger for the policy shift was a wave of AI-generated sexual images that used Grok to “undress” people in photos, including high profile targets. According to detailed accounts, Grok’s image-editing tools were used to create explicit deepfakes of UK Prime Minister Keir Starmer, turning a political leader into raw material for harassment and forcing regulators to confront how easily the system could be abused. Reporting on the fallout describes how the chatbot, which runs on Elon Musk’s platform X, allowed users to upload photos and request sexualized alterations that stripped away clothing or altered bodies in ways that would be impossible with conventional editing tools.
Under pressure from the UK government and other officials, X responded by restricting Grok’s image-editing and generation features to paying subscribers. The company placed the tools behind a subscription tier, effectively turning what had been a free capability into a premium feature, a change that was confirmed when X publicly stated that Grok image editing would now require payment. The move was framed as a safety measure that would limit abuse by tying powerful tools to verified, paying accounts, but it immediately raised questions about whether the platform was actually reducing harm or simply charging for access to the same risky functionality.
Critics say the paywall is “window dressing”
Political leaders in Ireland have been among the most outspoken critics of the new approach, arguing that a subscription barrier does nothing to address the underlying danger of non-consensual deepfakes. Irish ministers have publicly described the change as “window dressing” and “clearly” insufficient, warning that the Irish Government cannot take comfort from a model that still allows paying users to generate abusive content. One senior figure went so far as to deactivate their own X account in protest, underscoring how seriously officials view the risk of AI tools that can strip clothing from images or fabricate sexual scenes.
Those concerns are not limited to Ireland. Coverage of the backlash notes that the UK government’s pressure on X followed the circulation of AI-generated images of Keir Starmer, and that officials across Europe are now scrutinizing how Grok operates. In that context, the Irish criticism that the paywall is window dressing captures a broader fear: that platforms will respond to scandals with cosmetic tweaks rather than structural safeguards. By leaving the core capability intact for subscribers, X is seen as prioritizing revenue and user growth over the safety of people whose images can be manipulated without consent.
Safety by subscription, or monetizing harm?
From a product design perspective, X’s decision reflects a belief that tying powerful AI tools to paid accounts will deter the worst behavior. The logic is that subscribers are easier to identify, more likely to use their real identities and less inclined to risk losing access. Yet critics argue that this approach misunderstands how abuse works. People who are determined to create sexualized deepfakes of public figures or ex-partners are often willing to pay for the privilege, and the subscription model may simply filter out casual experimentation while leaving serious offenders untouched. In that sense, the paywall risks turning harmful capabilities into a premium service rather than a restricted one.
Investigations into how Grok behaves in practice reinforce those doubts. Tests using a free account found that image generation was blocked, but that did not mean the underlying model had been fundamentally changed or retrained. Instead, the functionality was simply unavailable unless the user upgraded, which is why critics say the platform has not fixed the “undressing” problem so much as moved it behind a credit card. One detailed examination of Grok’s behavior concluded that the system still posed serious risks, and that the shift to a paid tier just makes people for access to the same problematic tools.
Regulators push for real guardrails on Grok
Regulators are increasingly signaling that they expect more than access controls and subscription gates when it comes to AI systems that can generate or edit images. In the case of Grok, officials have focused on the need for technical guardrails that prevent the creation of non-consensual sexual content in the first place, regardless of whether the user is paying. That means robust filters that detect nudity, clothing removal and other forms of sexualization, as well as monitoring systems that can identify patterns of abuse and shut down offending accounts quickly. The Irish Government, which has already criticized the paywall, is pressing for stronger protections and has made clear that a subscription model does not satisfy its concerns.
International scrutiny is also growing. Detailed reporting on the controversy notes that Elon Musk’s AI bot has already had to limit image generation amid a wider backlash over deepfakes, particularly those targeting Keir Starmer and other public figures. In response, X has said that Grok’s image tools are being refined and that some capabilities have been temporarily restricted while the company works on better safeguards. Yet regulators point out that the platform only moved after intense public pressure, and that the current mix of partial limits and a paywall still leaves room for abuse. The fact that Grok had to be curtailed following a scandal, rather than designed with strict protections from the start, is now a central part of the political critique.
What the Grok backlash means for AI platforms
The controversy around Grok is bigger than one chatbot. It highlights a structural tension in the AI industry between rapid feature rollouts and the slow, difficult work of building safety into generative systems. Elon Musk has promoted Grok as a cutting edge assistant that can generate text and images with a distinctive personality, but the deepfake scandal shows how quickly those capabilities can be turned against people who never consented to be part of the experiment. When a tool can take a photo of Keir Starmer or an ordinary user and produce a sexualized fake in seconds, the line between innovation and abuse becomes dangerously thin.
Public reaction to the paywall also shows that users and regulators are no longer satisfied with surface level fixes. Critics argue that putting Grok’s image tools behind a subscription is not just inadequate but offensive, because it appears to monetize the very risks that triggered outrage in the first place. Detailed accounts of the backlash describe how Grok’s sexualized images and paywalled “fix” have put Musk’s platform under fire, with victims’ advocates and policymakers demanding real accountability rather than a new revenue stream. One report on Grok and its sexualized images notes that the change has been widely interpreted as a business decision dressed up as safety, a perception that could shape how future AI products are judged.
There are signs that the political fallout is already reshaping how some officials engage with X. Coverage of the Irish response describes how media and technology ministers have publicly criticized the platform, with Cillian Sherlock of the Press Association reporting that one Irish minister deactivated their account in direct response to the Grok controversy. That act of protest, detailed in a report on Irish ministers and their reaction to the paywall, sends a clear signal that cosmetic changes will not be enough to restore trust. For AI platforms that want to operate at global scale, the message is blunt: safety cannot be something users have to buy, and it cannot be an afterthought once the damage is already done.
As the debate continues, Grok has become a case study in what not to do when an AI product is implicated in abuse. Instead of a transparent overhaul of its safety systems, X opted for a subscription barrier that many see as pointless and offensive. The lesson for other companies is that when generative tools cross the line into facilitating sexual violence and harassment, the only credible response is to redesign the system so that such misuse is technically impossible or vanishingly rare. Anything less will be treated, in the words of the Irish Government, as little more than window dressing.
More From The Daily Overview
Grant Mercer covers market dynamics, business trends, and the economic forces driving growth across industries. His analysis connects macro movements with real-world implications for investors, entrepreneurs, and professionals. Through his work at The Daily Overview, Grant helps readers understand how markets function and where opportunities may emerge.
