Federal Judge Sides with Meta in Legal Dispute Over AI Training Using Copyrighted Books

On Wednesday, a federal judge ruled in favor of Meta in a lawsuit brought by 13 authors, including Sarah Silverman, who alleged that the company unlawfully utilized their copyrighted works to train its AI models.

Federal Judge Vince Chhabria delivered a summary judgment, resolving the case without a jury and determining that Meta’s training of AI models on the copyrighted texts fell within the “fair use” doctrine of copyright law, thus rendering it permissible.

This ruling comes on the heels of a similar decision just days earlier, where another federal judge ruled in favor of Anthropic in a related lawsuit. Collectively, these outcomes are viewed as a favorable turn for the tech industry, which has been beset by legal disputes from media firms regarding the appropriateness of utilizing copyrighted content for AI training.

However, these decisions do not represent the sweeping victories that certain corporations had hoped for; both judges highlighted that their verdicts were limited in scope.

Judge Chhabria pointed out that this ruling does not mean that all AI training using copyrighted materials is allowed, but indicates that the plaintiffs in this case “made the wrong arguments” and failed to furnish enough substantial evidence to support the right ones.

“This ruling does not imply that Meta’s use of copyrighted materials for training its language models is lawful,” Judge Chhabria remarked in his decision. He further added, “In cases involving uses like Meta’s, it seems that plaintiffs will often prevail, especially when the cases are supported by well-documented evidence regarding the market impact of the defendant’s usage.”

Judge Chhabria concluded that Meta’s use of copyrighted works in this instance was transformative—suggesting that the company’s AI models did not merely reproduce the authors’ writings.

Moreover, the plaintiffs were unable to prove that Meta’s reproduction of the books adversely affected the authors’ market, an essential factor in assessing copyright infringement.

“The plaintiffs did not provide any notable evidence regarding market dilution at all,” Judge Chhabria stated.

Both the successes for Anthropic and Meta involve training AI models on literary works; however, several lawsuits remain active against tech companies concerning their use of other copyrighted materials. For instance, The New York Times is suing OpenAI and Microsoft for employing news articles for AI model training, while Disney and Universal are taking legal action against Midjourney for its use of films and television shows.

Judge Chhabria noted in his ruling that the validity of fair use defenses highly depends on the specific circumstances of each case, and some industries may have stronger justifications for fair use than others.

“It seems that markets for certain types of works (like news articles) may be more vulnerable to indirect competition from AI-generated outputs,” Chhabria observed.