Partial Legal Victory for Meta: It Was Permitted to Use Copyrighted Works for Training Its Model

Partial Legal Victory for Meta: It Was Permitted to Use Copyrighted Works for Training Its Model
Source: Pexels - Pavel Danilyuka

A federal court in San Francisco has ruled that Meta legally used authors' works to train its AI models. Judge Vince Chhabria's decision determined that the Facebook parent company's practices are protected under the fair use doctrine, though the judge cautioned this doesn't automatically make all similar cases lawful. The millions of books, academic articles and comics used to train Meta's Llama AI model were copyright protected, including works by authors Ta-Nehisi Coates and Richard Kadrey.

This case is one of dozens of legal battles currently proceeding through American courts that attempt to define the legal boundaries of AI technology development. Judge Chhabria explained that his decision was not based on Meta's practices being generally lawful, but rather on the plaintiffs' failure to present appropriate arguments and sufficient evidence. Boies Schiller Flexner LLP, representing the authors, respectfully disagreed with the ruling given Meta's historically unprecedented pirating and emphasized that the ruling still holds that training AI on copyrighted works without permission generally violates the law.

The ruling marks the second victory within a week for technology companies, after another federal judge ruled in favour of Anthropic in a similar case. In his judgment, Judge Chhabria expressed concern about market dilution, referring to how AI systems can produce substantial amounts of content with minimal human effort, significantly undermining traditional creative processes and human creative motivation. Meta's case involved using LibGen, a shadow library that hosts much of its content without permission from rights holders, raising additional legal questions about copyright infringement.

Sources:

Meta AI Court Ruling Thumbnail
Federal Judge Sides with Meta in AI Copyright Lawsuit A federal court finds Meta's use of copyrighted books for AI training constitutes fair use, rejecting authors' claims in landmark case.
Meta wins AI copyright case in blow to authors
Court finds using online books to train artificial intelligence models without writers’ consent is ‘fair use’
Meta wins AI copyright lawsuit as US judge rules against authors
Writers accused Facebook owner of breach over its use of books without permission to train its AI system