In a Reuters interview on Feb 17, California Attorney General Rob Bonta, the state's top law enforcer, said his office sent a cease-and-desist to xAI on Jan 16 and is building an "AI oversight, accountability and regulation" program inside the Attorney General's office. Translation: California is not waiting for Washington, DC, to move first. Read more at Investing.com.
What triggered it:
Grok, xAI's conversational chatbot that is integrated with X (the social platform formerly known as Twitter), was used to generate non-consensual sexualized images of adults and potentially minors. Bonta says xAI deflected responsibility and still allowed some sexualized content for paying users. He added that stopping future behavior does not erase past harms. Bonta's office is seeking confirmation that the conduct has stopped and is in talks with the company. More details at Investing.com.
xAI's response:
In January, xAI said it added measures to reject requests for sexualized images of real people, for example editing outputs to show a bikini. The company said it blocks such generation where it would be illegal. At the same time, some features were placed behind a paywall, which regulators flagged as a problem. Read the report at The Star.
Regulatory backdrop:
This is not just a California issue. Ireland's Data Protection Commission (DPC) opened a probe under the General Data Protection Regulation (GDPR) into X over Grok's deepfakes. Spain has also taken steps on AI-generated child sexual abuse content linked to platforms. In short, regulators worldwide are paying attention. See the coverage at AP News.
Corporate structure:
xAI was acquired by Elon Musk's SpaceX on Feb 2, 2026. Yes, the rocket company. That acquisition ties xAI, X, Grok, and underlying AI infrastructure more tightly to SpaceX, which broadens the corporate footprint of any probe. Source: The Washington Post.
The policy play:
Bonta warned against leaving AI regulation only to Congress, citing federal gridlock. He pointed to a California bill, SB 69, that would require the Attorney General to build in-house AI expertise. His office also told OpenAI it has an ongoing interest after helping oversee OpenAI's 2025 recapitalization. Connecticut Attorney General William Tong called AI harms "the consumer protection fight of our time." Read more at The Star.
Why founders should care:
State enforcement is happening now. Key dates: cease-and-desist on Jan 16, 2026; Bonta's interview on Feb 17, 2026. Expect other state attorneys general to follow.
Safety cannot be a paywall exception. If premium tiers allow content that regulators prohibit, enforcement will follow. Parity across product tiers reduces legal risk.
Paper trails matter. Document safety tests, incident response steps, geoblocking logic, and model behavior changes. Keep records so you can show what you tested and when.
For more on enforcement timing and expectations, see Investing.com.
Short version:
Build safe models, verify the fix, document everything, and assume California will check your work. Probably more than once.
Sources:
Investing.com: California builds AI oversight unit and presses on xAI investigation - Investing.com
Investing.com: California AG sends letter demanding xAI stop producing deepfake content - Investing.com
The Star: California builds AI oversight unit and presses on xAI investigation - The Star
AP News: Ireland's DPC opens probe into X over Grok deepfakes - AP News
The Washington Post: SpaceX acquires xAI - Washington Post
Get daily insider tech news delivered to your inbox every weekday morning.
