California Attorney General Rob Bonta is advancing plans to establish a formal artificial intelligence accountability program within his office, as regulators examine xAI over the alleged generation of non-consensual sexually explicit images. Speaking in an interview, Bonta confirmed that the inquiry remains active and that his office is continuing discussions with the company.
According to Bonta, the California Attorney General’s office acted swiftly last month by issuing a cease-and-desist letter to xAI. The action came as regulators across multiple jurisdictions began reviewing the company’s AI chatbot, Grok, over reports that it produced sexualized content involving adults and potentially minors. He said his office is seeking assurances that the conduct has ceased entirely, while noting that xAI has disputed responsibility and continues to allow certain forms of sexualized content generation for paying subscribers.
“Just because you stop going forward doesn’t mean you get a pass on what you did,” Bonta added, underscoring the state’s stance on accountability.
The company, which was recently acquired by SpaceX, did not respond to requests for comment. In January, xAI stated that it had implemented safeguards to reject user prompts requesting sexualized images of real individuals, including modifying outputs such as editing subjects to appear in a bikini. The company has also said it restricts such image generation in jurisdictions where it is prohibited by law.
California’s actions signal a broader willingness by the state to position itself as a leading AI regulator, even as segments of the technology industry and some Republican lawmakers argue that oversight should rest primarily with federal authorities. Bonta cautioned against granting Congress exclusive regulatory control, pointing to past legislative gridlock around data privacy and artificial intelligence policy.
As part of this effort, the Attorney General’s office is “beefing up” its internal capabilities through what Bonta described as an “AI oversight, accountability and regulation program.” He emphasized that AI chatbots engaging in sexually explicit conversations with minors or providing instructions on self-harm are unacceptable and demand firm oversight.
State regulators have also communicated with OpenAI, headquartered in San Francisco, indicating California’s “ongoing interest” in ensuring the safety of its products and services. Bonta’s office was involved in overseeing aspects of OpenAI’s corporate restructuring last year, further reflecting the state’s active engagement with major AI developers.
Meanwhile, California lawmakers are considering legislation that would formally require the attorney general’s office to establish a dedicated program to deepen its AI expertise.
In a joint interview, William Tong, the Attorney General of Connecticut, described AI and social media-related harms as “the consumer protection fight of our time,” suggesting the issue could surpass the scale of past battles over opioids. “This affects all of our children,” he said, highlighting the growing urgency among state leaders to address emerging risks in the AI landscape.
