The federal judge allowed the AI copyright lawsuit against Meta to proceed, dismissing part of the complaint.
Authors Richard Kadrey, Sarah Silverman, and Ta-Nehisi Coates sued Meta for allegedly using their books to train the large language model (LLM) Llama without permission. They claimed that Meta intentionally removed copyright information to conceal the infringement. Although part of the complaint was dismissed, U.S. District Judge Vince Chhabria allowed the lawsuit to continue.
During last month's hearing, Judge Chhabria expressed skepticism about fully dismissing the case, although he also criticized the plaintiffs' lawyers' tone. Meta argued that its use of the authors' books falls under "fair use" under copyright law.
Meta accused of deleting copyright information
The Friday ruling affirmed that the copyright infringement allegations are sufficient to proceed with the lawsuit. More importantly, Judge Chhabria found that the authors have a plausible basis to allege that Meta intentionally deleted copyright management information (CMI) to conceal the infringement.
"These allegations, taken together, state a plausible, though not particularly strong, hypothesis that Meta deleted CMI to prevent Llama from generating outputs containing copyright information, thereby concealing its use of copyrighted materials in training," Chhabria wrote.
However, the judge dismissed the claims related to the California Comprehensive Computer Data Access and Fraud Act (CDAFA), as the authors failed to demonstrate that Meta directly accessed their computers or servers, but only used the book content.
The lawsuit has revealed a glimpse of Meta's copyright strategy. The plaintiffs' documents show that Mark Zuckerberg approved the use of copyrighted works to train Llama. Additionally, internal discussions at Meta indicate considerations about using content with unclear legal status for AI training.
This lawsuit is one of many copyright cases related to AI that are being reviewed by the courts, including The New York Times' lawsuit against OpenAI and Microsoft.
