California Orders Musk's xAI to Stop Grok's "Undressing" Feature After 3 Million Sexualized Images

California has ordered Elon Musk's artificial intelligence company to immediately stop what regulators describe as one of the largest nonconsensual imagery operations ever documented on a consumer AI product.

Attorney General Rob Bonta sent xAI a cease-and-desist letter demanding the company halt the creation and distribution of deepfake, nonconsensual intimate images. The letter cited "numerous examples of xAI taking ordinary, clothed images of women and children" and allowing users to depict them "in suggestive and sexually explicit scenarios and 'undress' them, all without the subjects' knowledge or consent."

The Scale of the Allegations

The scale alleged is staggering. A technical audit by the Center for Countering Digital Hate estimated that during an 11-day window between December 2025 and January 2026, Grok was used to generate more than 3 million sexualized images. Roughly 20,000 of those appeared to depict minors, according to the investigation.

"Spicy Mode" Under Fire

At the center of the case is Grok's so-called "Spicy Mode," which safety advocates argue was effectively "illegal by design" because of its nudification capability. Bonta's office gave xAI five days to demonstrate compliance and detail the steps it is taking to prevent the tool from "undressing" subjects in uploaded photographs.

Global Fallout

The fallout is now global. Japan has opened its own probe into X and Grok, Britain and Canada are moving forward with inquiries, and Malaysia and Indonesia have temporarily blocked access to Grok altogether.

xAI has not publicly detailed its compliance response. The case lands as Musk continues to navigate the aftermath of his falling-out with President Trump and mounting regulatory pressure on multiple continents.

Sources: TechCrunch, CalMatters