Mark Zuckerberg defends use of pirated AI books by using YouTube’s piracy handling: All the details

Mark Zuckerberg

Meta Platform’s CEO Mark Zuckerberg has again found himself in limelight as court documents reveal his defense strategy in an ongoing AI copyright case. The deposition, part of the Kadrey v. Meta lawsuit accuses Meta of using LibGen, a database that uses pirated e-books, to train its AI models, such as the Llama series. Zuckerberg compared Meta’s actions to YouTube’s efforts to remove pirated content, arguing that using such data sets is not unreasonable.

As mentioned in a report from TechCrunch, Meta has been accused of downloading materials from LibGen, with internal warnings coming from its AI research teams about potential litigation. Prominent writers including Ta-Nehisi Coates and Sarah Silverman have accused Meta of training their Llama AI models on data from a collection of pirated books.

However, during his deposition, Zuckerberg has defended the use of these books by stating: “For example, YouTube, I think, may end up hosting some stuff that people pirate for some time, but YouTube is trying to take that stuff down. And the vast majority of the stuff on YouTube, I would assume, is kind of good and they have the license to do.”

Moreover, when one of the plaintiffs' attorneys, David Boies, questioned the company's use of the LibGen dataset, Zuckerberg defended it by arguing that prohibiting the use of such resources for AI training would be unreasonable. Despite Zuckerberg’s remarks, court documents indicate that Meta staff members were unsure if utilizing LibGen was legal. Zuckerberg has also stated in the deposition that he knew very little about LibGen, even though there is proof that Meta had used it to train some AI models.