Mark Zuckerberg Gave Meta's Llama Team the OK To Train On Copyright Works, Filing Claims

The lawsuit, Kadrey v. Meta, alleges Meta used pirated copyrighted material to train its Llama AI models. Plaintiffs, including authors Sarah Silverman and Ta-Nehisi Coates, claim CEO Mark Zuckerberg authorized this action. Meta allegedly used the LibGen dataset, known for containing pirated ebooks and articles. Internal Meta communications reportedly revealed concerns about LibGen's legality and potential regulatory repercussions. Despite these concerns, Zuckerberg approved its use, according to the plaintiffs. The lawsuit claims Meta removed copyright information from the LibGen data to conceal its infringement. The New York Times previously reported on Meta's shortcuts in data acquisition for its AI. Meta's alleged actions involved using a dataset known to be pirated, potentially undermining its position with regulators. The plaintiffs argue Meta’s actions constitute copyright infringement.