Mark-Zuckerberg
A U.S. federal judge has ruled in favor of Meta in a copyright lawsuit brought by 13 authors, including comedian and writer Sarah Silverman, over the company’s use of their books to train AI models.
Judge Vince Chhabria granted summary judgment—meaning no jury trial was needed—finding that Meta’s use of the copyrighted works fell under the legal doctrine of “fair use.” In the judge’s view, the way Meta’s AI models used the books was considered “transformative,” not a direct reproduction.
This ruling follows a similar win for Anthropic just days earlier, strengthening Silicon Valley’s legal position in the debate over whether training AI models on copyrighted material is lawful. Still, the decisions aren’t sweeping victories for the tech industry.
Judge Chhabria emphasized that this specific outcome doesn’t mean all AI training with copyrighted material is legal. Instead, he stated the plaintiffs failed to provide strong arguments or compelling evidence, especially regarding market harm caused by Meta’s AI. “The plaintiffs presented no meaningful evidence on market dilution at all,” he noted.
He also made clear that future lawsuits—with better evidence and focused claims—could reach different outcomes. For instance, AI training using news content, films, or TV shows may present greater legal challenges. That’s especially relevant as lawsuits from The New York Times (against OpenAI and Microsoft) and from studios like Disney and Universal (against Midjourney) make their way through the courts.
The case reaffirms that “fair use” rulings depend heavily on context, with courts likely to evaluate each type of work—and how AI models use them—on a case-by-case basis.
About the Author
