On Wednesday, Meta received a legal victory when a federal judge, Vince Chhabria, dismissed a lawsuit from 13 authors, including Sarah Silverman. They had accused the company of unlawfully using their copyrighted books to train AI models.

Judge Chhabria gave his decision without a jury's involvement, ruling in Meta's favor. He concluded that using copyrighted books to teach AI models falls within the legal boundaries of the "fair use" copyright law principle.

This ruling closely succeeded a similar verdict in favor of Anthropic, steadily shifting recent legal battles toward the tech industry's favor. Yet, despite the victories, these cases did not provide the broad defense major tech firms were seeking, since both judgements were narrow in their scope.

Judge Chhabria clarified explicitly that his judgment should not be generalized to suggest that training AI models on copyrighted material is universally lawful. He underlined that the claimants in this particular case had presented weak arguments and weren't able to present strong evidence in favor of the serious ones.

Chhabria clarified, “In contexts similar to Meta's, it seems like the claimants will frequently win, provided those cases have more robust evidence on the impact on the marketplace due to the defendant's usage.”

To clarify, Chhabria stated that Meta's usage of copyrighted material in this instance was transformative, meaning that the AI models did more than just mimic the copyrighted books.

The plaintiffs couldn't sway the judge that Meta's utilization of the books negatively affected the authors' market, a fundamental aspect in determining a copyright infringement.

Judge Chhabria responded, “The plaintiffs haven't presented any significant proof regarding market dilution.”

Further victories for Anthropic and Meta in lawsuits related to training AI models on books, don't cover other pending lawsuits against tech firms accused of training AI models on other copyrighted assets. For example, OpenAI and Microsoft are being sued by The New York Times for using news articles, and Midjourney is battling a lawsuit from Disney and Universal for its use of films and TV shows.

Judge Chhabria pointed out during his verdict that a fair use defense heavily relies on the specifics of the case, adding that some industries might have a stronger claim to fair use than others.

In Chhabria’s own words, “It appears that markets for certain kinds of works, such as news articles, could be more exposed to indirect competition from AI outputs.”