In a significant legal victory for Meta, the tech giant emerged triumphant in an AI copyright case filed by 13 authors who accused the firm of improperly using their writings to train its artificial intelligence systems without consent. The judge presiding over the case was Vince Chhabria. ruling in Meta’s favor, stating it is entitled to summary judgment regarding its fair use defense against the allegation that using copies of these plaintiffs' books as training data for large language models constituted an infringement.
Nevertheless, the judge highlighted certain vulnerabilities within the ecosystem of Big Tech’s AI initiatives and Meta’s justifications for their actions being considered fair use. "This decision doesn’t imply that Meta’s utilization of copyrighted material to develop its language models is legal," stated Judge Chhabria.
"It simply indicates that these plaintiffs presented incorrect arguments and did not build a sufficient case for the correct argument." This decision comes after Anthropics significant win regarding fair use, which they achieved when another federal judge ruled earlier that using their models to train on legally acquired book copies was permissible. is fair use .
Judge Chhabria indicates that out of the authors' arguments concerning fair use, two were "clearly weak": firstly, the capability of Meta’s Llama AI to replicate excerpts from their books; secondly, Meta utilizing their work to develop its AI models without authorization, which diminishes their capacity to sell licenses for such usage. The judge stated, "The amount of text generated by Llama from the plaintiffs' books is insignificant, and they do not have the right to control the market for licensing their works as AI training material."
According to Judge Chhabria, the plaintiffs failed to present a "potentially winning argument" that Meta's replication efforts would result in "products that may inundate the market with comparable offerings, leading to market dilution." Additionally, he referenced the Anthropic decision, noting that Judge William Alsup dismissed worries regarding potential damage generative AI might cause to the market for the training materials used.