Elon Musk's company X has announced significant new restrictions on its artificial intelligence chatbot, Grok, preventing it from editing photographs to undress real people. The move comes after a fierce international backlash and multiple government investigations into the proliferation of sexually explicit deepfake material.
Global Pressure Forces Policy Change
The announcement was made in a statement posted on the X platform late on Wednesday, 14 January 2026. It stated that technological measures are now in place to stop the Grok account from allowing the editing of images of real people in revealing clothing. This specifically includes attire such as bikinis, underwear, and other similar garments.
The decision followed what the company described as a "global backlash" over sexualised images of women and children generated using the tool. Initially, media enquiries on the issue were met with the dismissive response: "legacy media lies." However, the mounting pressure from regulators and the public has prompted a stark reversal in policy.
Investigations and International Bans
The pushback has been swift and widespread. Notably, the state of California announced an investigation on Wednesday into the spread of nonconsensual sexually explicit material produced via Grok. California's Attorney General, Rob Bonta, was cited saying the "avalanche of reports" was "shocking" and that the material had been used to harass people across the internet.
Several countries have already taken legal action. Malaysia and Indonesia have blocked access to Grok entirely. The United Kingdom and the European Union have launched probes into potential breaches of online safety laws. Furthermore, France, India, and Brazil have issued formal warnings or called for investigations into the AI's misuse.
New Safeguards and Accountable Access
In its statement, X's AI division, xAI, clarified that the new rule applies universally to all users, including those with paid subscriptions who typically enjoy broader feature access. The company emphasised its stance, stating: "We have zero tolerance for the AI-based creation and dissemination of nonconsensual intimate images or of child sexual abuse material."
To bolster accountability, xAI has also limited all image creation and editing functions to paid subscribers only. This measure is intended to ensure that individuals who attempt to abuse Grok to violate the law or platform policies can be more easily identified and held responsible.
The controversy centred largely on Grok's "spicy mode," which users had exploited to generate explicit content. The company now says it will employ geoblocking to prevent the tool from operating in ways that violate specific local laws, marking a significant step towards tighter regional control of its AI's capabilities.
