In a recent legal dispute, a federal judge in San Francisco decided in favor of Meta Platforms, the parent company of Facebook, by dismissing a copyright infringement lawsuit brought by a group of authors. These authors had claimed that Meta unlawfully used their works to enhance its artificial intelligence systems. The decision from U.S. District Judge Vince Chhabria marked the second dismissal of a major lawsuit that week concerning copyright issues related to AI development.
Judge Chhabria concluded that the 13 authors involved in the case presented flawed arguments, leading to the dismissal. However, he clarified that the decision is restricted to this particular case and does not imply a blanket legality of Meta’s practices regarding copyrighted materials. In his ruling, Chhabria stated that the authors failed to make the correct legal points to substantiate their claims.
The plaintiffs, which include high-profile figures such as comedian Sarah Silverman and esteemed authors Jacqueline Woodson and Ta-Nehisi Coates, expressed their disappointment in a statement. They argued that the court recognized the violation of copyright law by companies that use protected works without permission or payment, yet still favored Meta. Meta expressed satisfaction with the ruling, highlighting the importance of fair use in fostering innovation and creativity with open-source AI models.
Despite Meta’s success in this legal battle, the outcome may be seen as a tentative win. In the 40-page ruling, Judge Chhabria indicated skepticism regarding Meta’s and other AI companies’ adherence to copyright law, hinting at the potential for future lawsuits if authors better frame their arguments. He dismissed the suggestion that copyright laws would impede technological progress, emphasizing the significant financial gains expected from AI advancements and suggesting compensation for copyright holders could be feasible.
In a related ruling earlier that week, U.S. District Judge William Alsup addressed a similar case against AI company Anthropic. He found that while training their chatbot on copyrighted books was legal under fair use, the method of acquiring those books from pirated sources warranted a trial. Alsup noted that AI technology’s transformative nature justifies its use under copyright laws.
The authors’ lawsuit against Meta centered on claims of widespread copyright infringement via the integration of their works in Meta’s AI system, Llama. They accused Meta of extracting text from pirated sources instead of securing legitimate access through purchase or licensing. Meta contended its use did not infringe on the original works since its AI-generated outputs differed significantly from the source material.
Meta defended its training methods by asserting that its system wouldn’t recreate entire works upon request, thus not hindering the consumption of the original texts. Despite allegations stemming from Meta’s utilization of “shadow libraries,” the company argued that its actions were no different than if they had collaborated with legitimate libraries. Historical parallels were drawn to Google’s decade-long legal tug-of-war, ultimately vindicated in 2016 concerning its Google Books project.
This legal challenge exposed internal deliberations within Meta, including top executives like CEO Mark Zuckerberg, concerning ethical considerations of leveraging pirated content databases. While the detailed outcome remains specific to the lawsuit by these 13 plaintiffs, Judge Chhabria remarked that the broader implications are minimal, not extending to other authors potentially affected by Meta’s AI training practices.