Meta on Wednesday prevailed against a group of 13 authors in a major copyright case involving the company’s Llama artificial intelligence model, but the judge made clear his ruling was limited to this case.
U.S. District Judge Vince Chhabria sided with Meta’s argument that the company’s use of books to train its large language models, or LLMs, is protected under the fair use doctrine of U.S. copyright law.
Lawyers representing the plaintiffs, including Sarah Silverman and Ta-Nehisi Coates, alleged that Meta violated the nation’s copyright law because the company did not seek permission from the authors to use their books for the company’s AI model, among other claims.
Notably, Chhabria said that it “is generally illegal to copy protected works without permission,” but in this case, the plaintiffs failed to present a compelling argument that Meta’s use of books to train Llama caused “market harm.” Chhabria wrote that the plaintiffs had put forward two flawed arguments for their case.
“On this record Meta has defeated the plaintiffs’ half-hearted argument that its copying causes or threatens significant market harm,” Chhabria said. “That conclusion may be in significant tension with reality.”
Meta’s practice of “copying the work for a transformative purpose” is protected by the fair use doctrine, the judge wrote.
“We appreciate today’s decision from the Court,” a Meta spokesperson said in a statement. “Open-source AI models are powering transformative innovations, productivity and creativity for individuals and companies, and fair use of copyright material is a vital legal framework for building this transformative technology.”
Though there could be valid arguments that Meta’s data training practice negatively impacts the book market, the plaintiffs did not adequately make their case, the judge wrote.
Attorneys representing the plaintiffs said in a statement said that they “respectfully disagree” with the decision.
“The court ruled that AI companies that ‘feed copyright-protected works into their models without getting permission from the copyright holders or paying for them’ are generally violating the law,” the statement said. “Yet, despite the undisputed record of Meta’s historically unprecedented pirating of copyrighted works, the court ruled in Meta’s favor.”
Still, Chhabria noted several flaws in Meta’s defense, including the notion that the “public interest” would be “badly disserved” if the company and other businesses were prohibited “from using copyrighted text as training data without paying to do so.”
“Meta seems to imply that such a ruling would stop the development of LLMs and other generative AI technologies in its tracks,” Chhabria wrote. “This is nonsense.”
The judge left the door open for other authors to bring similar AI-related copyright lawsuits against Meta, saying that “in the grand scheme of things, the consequences of this ruling are limited.”
“This is not a class action, so the ruling only affects the rights of these thirteen authors — not the countless others whose works Meta used to train its models,” he wrote. “And, as should now be clear, this ruling does not stand for the proposition that Meta’s use of copyrighted materials to train its language models is lawful.”
Additionally, Chhabria noted that there is still a pending, separate claim made by the plaintiffs alleging that Meta “may have illegally distributed their works (via torrenting).”
Earlier this week, a federal judge ruled that Anthropic’s use of books to train its AI model Claude was also “transformative,” thus satisfying the fair use doctrine. Still, that judge said that Anthropic must face a trial over allegations that it downloaded millions of pirated books to train its AI systems.”
“That Anthropic later bought a copy of a book it earlier stole off the internet will not absolve it of liability for the theft, but it may affect the extent of statutory damages,” the judge wrote.
