Social media platform X has announced new restrictions on its artificial intelligence tool Grok, confirming that it will no longer allow the editing of images of real people to show them in revealing clothing in jurisdictions where such content is illegal. The move follows widespread backlash over the creation and circulation of sexualised AI deepfakes, particularly targeting women.
The announcement comes after mounting pressure from regulators, governments, campaigners, and victims who say the harm caused by the technology has already been significant.
What X Has Changed About Grok
X said it has implemented “technological measures” to block Grok from editing images of real people into sexualised or revealing forms, including bikinis, underwear, and similar attire, in countries where the practice breaches the law.
The platform stated that it has introduced geoblocking to limit the feature based on local laws. It also reiterated that only paid users can access image-editing tools on Grok, which it says adds an extra layer of accountability.
Grok, which Elon Musk launched on X in 2023, previously allowed users to generate or edit images that many critics described as sexualised deepfakes. These images often appeared publicly on the platform, making them widely accessible.
UK Government and Regulator React
The UK government welcomed the move, with officials describing it as a vindication of earlier calls for X to tighten controls around Grok.
The communications regulator Ofcom called the change a “welcome development” but confirmed that its investigation into whether X breached UK law remains ongoing. Ofcom said it is still seeking answers about how the abuse occurred and what safeguards X failed to put in place initially.
Technology Secretary Liz Kendall welcomed the announcement but stressed that Ofcom must fully establish the facts through its investigation.
Victims Say the Action Came Too Late
Campaigners and victims argue that the platform acted only after serious damage had already occurred. Journalist and campaigner Jess Davies, whose images were edited using Grok, described the changes as a positive step but said the harm should never have happened.
She said the public nature of X made the abuse especially damaging, as the images were easy to find, share, and weaponise. While she acknowledged the policy shift, she criticised X’s broader response as inadequate and slow.
Davies warned that many women have already suffered reputational damage, emotional distress, and fear for their personal safety as a result of AI image abuse.
Academics and Campaigners Welcome the Shift, With Reservations
Dr Daisy Dixon of Cardiff University, who also experienced AI image abuse using Grok, described the platform’s reversal as a “battle-win” for campaigners. However, she said it does not erase the impact on those affected.
She noted that such abuse alters how women experience their bodies and public presence online, often without consent or recourse. According to Dr Dixon, platforms must take responsibility before harm occurs, not after public outrage.
Andrea Simon, director of the End Violence Against Women Coalition, said the decision showed that pressure from victims, activists, and governments can force tech companies to act. She added that platforms must now adopt proactive measures to prevent future AI-generated abuse.
Political Pressure Mounts on X
X announced the changes just hours after California’s top prosecutor confirmed that the state was investigating the spread of sexualised AI deepfakes, including content involving children. The probe added to growing international scrutiny of AI image tools.
Earlier, Musk had defended X by arguing that critics wanted to suppress free speech. He also shared AI-generated images of UK Prime Minister Keir Starmer in a bikini, which drew further criticism from politicians and regulators.
Starmer later warned that X could lose the “right to self-regulate” if it failed to act responsibly. He said the government would strengthen legislation if platforms continued to allow harmful AI content.
Enforcement and Geoblocking Concerns
Despite X’s announcement, questions remain about how the company will enforce the new rules. Experts say it remains unclear how Grok will reliably identify whether an image depicts a real person or how the system will respond when users attempt to bypass restrictions.
Policy researcher Riana Pfefferkorn said she was surprised X did not disable the editing features as soon as abuse emerged. She also questioned whether location-based restrictions would be effective.
Users often use tools such as virtual private networks to bypass geoblocks. VPN usage increased in the UK last year after age verification rules came into force under the Online Safety Act, raising concerns that similar tactics could undermine Grok’s new safeguards.
Ofcom Investigation Continues
Ofcom confirmed it is investigating whether X failed to comply with UK law regarding the distribution of sexualised AI images. The regulator has the power to seek court orders requiring internet service providers to block access to platforms that repeatedly breach the law.
The regulator said X must demonstrate meaningful compliance, not just policy announcements. If the platform fails to meet its legal obligations, further enforcement action may follow.
A Broader Reckoning for AI Platforms
The Grok controversy highlights wider concerns about AI-generated abuse and the responsibilities of tech companies. Critics argue that platforms often prioritise innovation and engagement over safety, leaving victims to deal with the consequences.
While X’s move marks a shift in approach, campaigners say it must form part of a broader effort to regulate AI tools, protect users, and ensure accountability. As governments tighten scrutiny, the way platforms deploy generative AI may soon face stricter legal limits worldwide.
The coming months will test whether X’s changes prevent further abuse or whether regulators will need to step in more forcefully to protect users from AI-driven harm.
