On March 5, 2026 the court issued its order in x.AI v. Bonta, denying x.AI a preliminary injunction. x.AI asked the court to stay enforcement of the requirement in AB 2013, entitled “Artificial Intelligence Training Data Transparency,” that requires developers of “a generative artificial intelligence system or service” that is “publicly available to Californians for use” to “post on the developer’s internet website documentation regarding the data used by the developer to train the generative artificial intelligence system or service.” x.AI had complied, but questioned whether its disclosure satisfied their statutory obligation.
Standing
At oral argument on February 23, 2026, the court asked the State "whether there is any current intention to institute an enforcement action against Plaintiff." The State responded that it "cannot comment on whether or when the Attorney General intends to pursue specific enforcement actions."
For standing for a pre-enforcement claim, "a plaintiff must … allege (1) an intention to engage in a course of conduct arguably affected with a constitutional interest, but (2) proscribed by a statute, and (3) there exists a credible threat of prosecution thereunder." The first two prongs were satisfied and the State's failure to disavow enforcement was enough to establish a credible threat of enforcement. x.AI therefore has standing to bring the case.
Likelihood of Success on the Merits
Fifth Amendment Takings (trade secrets)
x.AI alleged that the duty to publicly disclose information about training data was a taking of its trade secrets. Trade secrets can be subject matter of a taking, but x.AI failed to provide sufficient information for the court to establish that x.AI had a trade secret in information that it had to disclose: "Plaintiff’s Complaint trades in frequent abstraction and hypotheticals, rather than pleading specifics about Plaintiff’s practices. … Plaintiff’s resort to generalizations and hypotheticals about the AI model development industry make it difficult for the Court to find that Plaintiff has carried the heavy burden of showing a likelihood of success in proving that trade secrets are at play here." x.AI therefore failed to show that it is likely to succeed on the merits in its takings claim.
First Amendment (compelled speech)
x.AI also alleged that the obligation to provide the required information about training data was unlawful compelled speech under the First Amendment. The court agreed that the law was a content-based restriction on speech, implicating the First Amendment, but decided that this was commercial information subject to an intermediate degree of scrutiny under Central Hudson Gas & Elec. Corp. v. Pub. Serv. Comm'n of New York, 447 U.S. 557, 561-62 (1980). Under intermediate scrutiny, "the State must establish that the law directly advance[s] a substantial governmental interest, and [that] the means chosen [are] not … more extensive than necessary." xAI claimed that the information was of no value, but the court disagreed. The information may be useful to consumers, who can review information about the datasets used to train the models, which was a government interest. Perhaps after a full review on the merits the disclosure may be more extensive than necessary, but for now "Plaintiff has demonstrated a distinct possibility of prevailing on the merits under Central Hudson. But it had not demonstrated a likelihood of success on the merits."
Fourteenth Amendment (vagueness)
Finally, xAI also challenged the statute for vagueness. The statute has a detailed list of the information that must be provided. The court wasn't buying x.AI's argument that the terms "datasets" and "data points" were ambiguous, since the plaintiff understood the meaning of these terms in other parts of the complaint. The fact that the list was open-ended also wasn't enough for vagueness: "given that a statute entirely lacking a list of factors can still be sufficiently clear, it is likely that a non-exhaustive list is enough."
The motion for a preliminary injunction was therefore denied.