Elon Musk’s artificial intelligence company xAI was unable to convince a California federal court on Thursday to temporarily block the state’s law requiring companies to disclose information about the data they use to train AI models. U.S. District Judge Jesus Bernal said that xAI had not yet shown it was likely to prove the law violated its free-speech rights or was, otherwise, unconstitutional.
A spokesperson for the California Department of Justice said the department “celebrates this key win and remains committed to continuing our defense” of the law.
California has a data transparency law — which was enacted by governor Gavin Newsom —that requires generative AI companies to publicly post a summary of the datasets used to train their systems. This law came to effect on Jan. 1, as part of a push for more AI regulations.
READ: Tesla invests $2 billion in Elon Musk’s xAI (January 29, 2026)
xAI sued the state in December 2025, arguing that the law violated its free-speech rights under the U.S. Constitution and would force the company to reveal trade secrets about how its AI models are trained. On Thursday, Bernal denied xAI’s request for a preliminary injunction to halt the law’s enforcement, finding the company had not shown at this stage in the case that its lawsuit was likely to succeed.
Meanwhile, the California Attorney General Rob Bonta is building an artificial intelligence accountability program to strengthen AI oversight amid federal regulatory gridlock. According to Reuters, Bonta’s office is also probing xAI over the generation of non-consensual sexually explicit images.
READ: Back-to-back founder exits rock Elon Musk’s xAI (
The office had moved quickly last month to send a cease-and-desist letter to xAI, as regulators globally investigated the company over sexualized content that its AI chatbot Grok produced of adults and potentially minors, Bonta said. Bonta said his office is seeking confirmation that the conduct has stopped and remains in discussions with the company. He said xAI deflected responsibility and still permits some 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.
The California Attorney General’s office is “beefing up” its in-house expertise through its “AI oversight, accountability and regulation program,” Bonta said. AI chatbots that have sexually explicit conversations with youth or tell them how to commit suicide are unacceptable, he said. The state’s legislature is considering a bill that would require the attorney general’s office to establish a program to build expertise in AI.