California AI Transparency Law Upheld by Federal Court

California’s Generative AI Training Data Transparency law came into effect on January 1. The law requires developers of generative AI systems to publish certain information on their websites. This information includes high-level summaries of the datasets used to train gen AI systems. X.AI challenged the law, arguing that its datasets are proprietary and protected as trade secrets. Last week, a federal judge in the District Court for the Central District of California disagreed, denying X.AI’s request for a preliminary injunction.  A recent Fisher Phillips memo discusses the Judge’s ruling examining X.AI’s three main arguments:

• “On trade secrets, Judge Bernal acknowledged that training datasets could potentially be trade secrets, but found X.AI’s “generalized, abstract pleading” failed to demonstrate its datasets are distinct from competitors’ in a way that merits protection. The court said X.AI’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.”
• On vagueness, the court rejected X.AI’s argument that terms like “dataset” and “data point” are undefined in the statute, noting the company “seems to understand and use with ease ‘dataset’ throughout its Complaint.”
• On free speech, Judge Bernal ruled X.AI had not shown the law violates First Amendment rights at this stage in the case.”

For now, the law will remain in effect. We’ll see if X.AI or other California-based AI companies bring further challenges. The memo notes that companies using third-party AI systems subject to the California law should ensure vendor compliance to avoid contractual or reputational harm.