A federal judge has allowed an AI copyright lawsuit against Meta to move forward, keeping alive central claims from authors who say the company used their books to train its Llama AI models without violating their intellectual property rights.
The case, Kadrey vs. Meta, remains only partly intact. U.S. District Judge Vince Chhabria dismissed one section of the lawsuit, but rejected Meta’s effort to end the case at this stage.
What the authors claim
The lawsuit was brought by authors including Richard Kadrey, Sarah Silverman, and Ta-Nehisi Coates. Their core allegation is that Meta used their books to train its Llama AI models, and that this use violated their intellectual property rights.
The authors also claim Meta removed copyright information from their books. Their argument is that the company did this to hide the alleged infringement connected to AI training.
That second point matters because it moves the dispute beyond the basic question of whether copyrighted books were used. It also asks whether Meta intentionally removed copyright management information, often shortened to CMI, in a way that helped conceal the alleged conduct.
In Friday’s ruling, Chhabria said the authors had adequately alleged that Meta intentionally removed CMI to conceal copyright infringement. He also wrote that the copyright infringement allegation is “obviously a concrete injury sufficient for standing.”
“Taken together, these allegations raise a ‘reasonable, if not particularly strong inference’ that Meta removed CMI to try to prevent Llama from outputting CMI and thus revealing it was trained on copyrighted material,” Chhabria wrote.
Why Meta’s dismissal argument did not fully work
Meta has argued that its training of AI models qualifies as fair use. It also said the case should be dismissed because the authors lack standing to sue.
The court did not accept dismissal of the main copyright claims at this point. Chhabria’s ruling found that the alleged copyright infringement was enough to establish a concrete injury for standing.
That does not decide whether Meta ultimately wins or loses on fair use. It means the authors’ claims can continue, and the court will allow the case to proceed rather than ending it now.
The ruling also fits with what Chhabria had suggested in court last month. At that hearing, he seemed to indicate he was against dismissal, even as he criticized what he viewed as “over-the-top” rhetoric from the authors’ legal teams.
The claim the judge dismissed
The authors did not keep every claim. Chhabria dismissed their claims related to the California Comprehensive Computer Data Access and Fraud Act, known as CDAFA.
The reason was narrow. According to the ruling described in the source article, the authors did not “allege that Meta accessed their computers or servers — only their data (in the form of their books).”
That distinction shaped the court’s decision on this part of the case. The authors alleged access to their data, but the judge found that was not the same as alleging access to their computers or servers for the purpose of the CDAFA claim.
So the lawsuit now moves forward without that part of the complaint. The surviving dispute is centered on AI copyright, Llama training, fair use, standing, and copyright management information.
What the filings have already revealed
The lawsuit has also offered glimpses into how Meta approached copyright questions around AI training, according to court filings from the plaintiffs.
Those filings claim that Mark Zuckerberg gave the Llama team permission to train the models using copyrighted works. They also claim that other Meta team members discussed the use of legally questionable content for AI training.
These claims are part of the broader factual backdrop of the case. They do not resolve the legal questions, but they show why the lawsuit has drawn attention beyond the individual authors involved.
For readers following AI and copyright, the case is important because it focuses on a central tension in generative AI: whether copyrighted material can be used to train models, and under what legal theory that use might be defended.
A wider AI copyright moment
Kadrey vs. Meta is not the only AI copyright lawsuit currently before the courts. The source article notes that courts are weighing a number of AI copyright lawsuits at the moment, including The New York Times’ lawsuit against OpenAI.
That wider legal context makes the Meta case more than a single dispute between authors and a technology company. It is part of a broader court process that is testing how existing copyright claims apply to AI model training.
For now, the practical result is clear: Meta has not escaped the lawsuit. The authors’ core claims over Llama AI models, copyrighted books, and removed copyright information remain alive, while the CDAFA portion has been dismissed.
The next stages will determine how far those surviving claims can go. But the ruling already sets an important marker: at least some author claims over AI training and copyright management information are strong enough to proceed in federal court.