A federal judge has indicated that Meta Platforms’ “fair use” defense in a closely watched AI copyright lawsuit might depend more on whether its AI models economically damage authors than on the company’s admitted use of pirated books for training.
During a May 1 hearing in the Northern District of California for the Kadrey v. Meta case, Judge Vince Chhabria suggested the central question is whether Meta’s Llama AI models harm the market for original works created by plaintiffs like authors Richard Kadrey, Sarah Silverman, and Ta-Nehisi Coates.
Market Harm vs. Piracy
The hearing, focused on motions for partial summary judgment, saw Judge Chhabria steer the discussion away from Meta’s sourcing of training data from “shadow libraries” – online sources like LibGen, Z-Library, Bibliotik, and Anna’s Archive known for hosting unauthorized content.
While calling Meta’s methods “kind of messed up,” the judge emphasized the potential market impact. Addressing Meta’s lawyer, Kannon Shanmugam, Judge Chhabria stated, “If you are dramatically changing, you might even say obliterating, the market for that person’s work, and you’re saying that you don’t even have to pay a license to that person to use their work to create the product that’s destroying the market for their work—I just don’t understand how that can be fair use.”
Shanmugam described potential market effects as “just speculation.” Fair use, a complex legal concept, allows limited use of copyrighted material without permission under specific circumstances, evaluated based on factors including the purpose of the use and its effect on the original work’s market value.
The ‘Next Taylor Swift’ Problem
Judge Chhabria used a Taylor Swift analogy to explore the potential harm, particularly to less established creators. He questioned the fairness of an AI trained on an artist’s music generating billions of similar songs. “What about the next Taylor Swift?” he asked, suggesting an emerging artist’s career could be hampered if AI models trained on their unique style saturated the market.
However, the judge also expressed strong doubts about the plaintiffs’ current ability to prove such harm. Represented by attorney David Boies, the authors were pressed on their evidence. “It seems like you’re asking me to speculate that the market for Sarah Silverman’s memoir will be affected,” the judge told Boies. “It’s not obvious to me that is the case.”
He reiterated that the legal standard wasn’t whether Meta’s actions were merely problematic, but whether they constituted copyright infringement: “The question, as the courts tell us over and over again, is not whether something is messed up but whether it’s copyright infringement.”
Lawsuit Background and Allegations
The case began in July 2023 when Silverman and other authors sued Meta (and initially OpenAI), alleging their books were used without consent to train AI models like Meta’s Llama. Large Language Models (LLMs) like Llama learn patterns and generate text by processing enormous datasets.
Court documents revealed in January 2025 contained allegations, based on internal communications, that Meta CEO Mark Zuckerberg approved using the LibGen dataset despite employee concerns. One Meta engineer wrote, “Torrenting from a [Meta-owned] corporate laptop doesn’t feel right.”
The filings also accused Meta of using BitTorrent protocols and deliberately stripping Copyright Management Information (CMI), such as author and copyright details, potentially violating the Digital Millennium Copyright Act (DMCA) and California’s Computer Data Access and Fraud Act (CDAFA).
Judge Chhabria had previously criticized Meta’s attempts to seal related documents as being “designed to avoid negative publicity.” Furthermore, reports in March 2025 suggested Meta may have reuploaded a substantial portion of the torrented data, potentially exposing it to separate claims of illegal distribution.
Meta’s Defense and Broader AI Copyright Battles
Meta argues its AI training methods are transformative and protected by fair use, a defense echoed by other tech giants in similar disputes, such as Microsoft and OpenAI against The New York Times.
The Kadrey case is part of a wave of litigation testing the application of copyright law to generative AI. The legal challenges are international, with French publishers filing suit in Paris, describing Meta’s alleged actions as “monumental looting.”
Regulatory bodies like the European Commission are also examining these practices under new rules like the EU AI Act. The outcome of the Kadrey case could set a meaningful precedent. Acknowledging the case’s significance and complexity, Judge Chhabria concluded the hearing without giving a decision date, joking, “Just kidding! I will take a lot longer to think about it.”
