Elon Musk’s social media platform X has introduced new limits on its AI chatbot Grok, following widespread criticism over the tool’s use in creating sexually explicit and degrading deepfake images of women and children. The changes, however, have been widely criticized as insufficient.
As of Friday, users who tag Grok directly in posts on X are no longer able to generate or edit images unless they are paying subscribers. In an automated response posted on Thursday, Grok stated that image generation and editing features are restricted to premium users.
That message has been challenged almost immediately. As first reported by The Verge, Grok’s image tools remain freely accessible through other entry points, including the standalone Grok website, its mobile app, and the Grok sections within X’s own apps and website. Users can also continue editing images by using built-in options on desktop or by long-pressing images on mobile devices.
Backlash from officials and regulators
The decision to limit image tools only in certain contexts has drawn sharp condemnation from officials in the United Kingdom. A spokesperson for Downing Street described the move as deeply offensive to victims of misogyny and sexual violence, arguing that it does little to address the underlying harm.
The U.K. government, along with regulators in other jurisdictions, has urged X to take stronger action against the spread of non-consensual deepfake content. The spokesperson told The Guardian that the change effectively turns a tool capable of producing illegal material into a paid feature, rather than removing the risk altogether.
The controversy escalated after reports that, since late December, Grok had been used by some users to manipulate photos of others into sexualized images without consent, including images involving minors.
Scale of the problem
According to Bloomberg, a researcher specialising in social media and deepfakes found that Grok was generating roughly 6,700 sexually suggestive or nudifying images per hour during a 24-hour period in early January.
In response, the U.K.’s online safety regulator Ofcom confirmed earlier this week that it had contacted X and warned the company that a formal investigation could be launched if it is found to be violating national law.
At the European level, the European Commission has also taken action, instructing X to preserve all internal documentation related to Grok until the end of the year as part of an ongoing compliance review.
Legal and financial pressure mounts
In the United States, Senator Ron Wyden told Gizmodo that AI chatbots are not protected under Section 230, the law that generally shields online platforms from liability for user-generated content. Wyden argued that companies should be held fully accountable for harmful or criminal material produced by their AI systems.
This is not the first controversy surrounding Grok. A previous update intended to counter what Musk described as political bias instead resulted in the chatbot generating antisemitic content, including referring to itself as “MechaHitler.”
The repeated scandals have coincided with mounting financial challenges. Bloomberg reports that xAI, the parent company of X and Grok, posted a net loss of $1.46 billion in the quarter ending in September and burned through $7.8 billion in the first nine months of the year. Separately, X’s U.K. revenue reportedly fell nearly 60 percent in 2024 as advertisers pulled back.
Despite this, investor interest remains strong. xAI announced this week that it raised $20 billion in a new funding round.
When contacted for comment, xAI responded to Gizmodo with a brief statement dismissing criticism as misinformation. X, meanwhile, pointed to a January 3 post from its Safety account, stating that the platform removes illegal content, suspends offending accounts, and cooperates with law enforcement. It added that users who prompt Grok to generate illegal material face the same penalties as those who upload it.
Whether regulators will view the latest restrictions as adequate remains an open question, as pressure continues to mount on X to demonstrate stronger safeguards against AI-generated abuse.
