Meta achieved a legal victory in a copyright lawsuit filed by 13 authors. The lawsuit claimed that Meta used their books without permission to train its artificial intelligence models, particularly LLaMA. The central issue in the case was whether Meta's practice of copying copyrighted books to train large language models (LLMs) violated U.S. copyright law. Ultimately, the judge ruled that Meta's actions fell under fair use and decided in favor of the company.
Judge Vince Chhabria stated that Meta was entitled to summary judgment on its fair use defense. The judge ruled that the plaintiffs did not present compelling legal arguments and failed to provide evidence to support their invoices.
The court was careful to point out that this ruling does not mean training AI on copyrighted materials is, in all cases, legal. Instead, the judge indicated that those legal arguments were not the proper legal basis in this specific case.
According to the plaintiffs, Meta's AIs, LLaMA included, reproduce passages from their books. Hence, it was said that Meta's conduct interfered with the authors' licensing of their works for AI training purposes. However, the judge rejected the claims.
The judge, finding that the actual range of generated output was too limited to sustain the alleged harm, also added that, in any event, there exists no guaranteed right for an author to license works for AI training.
Indeed, the court determined that the plaintiffs' claims of market dilution were without merit. Judge Chhabria determined that Meta's AI systems are not likely to saturate the market with similar works. In addition, the judge explained that the authors had not proved any actual market threat to the licensing market or future economic loss from Meta's actions.
A favorable court decision highlights the significance of presenting strong legal arguments in copyright cases involving AI. While verdicts do not establish widely enforced precedents, they serve as reminders of the obstacles that poorly conceived lawsuits face. As AI technology advances, clearer legal standards regarding the use of copyrighted training data will emerge to address these issues in the context of ongoing AI development.