Meta Ordered to Name Executives Who Authorized Copyrighted AI Training Data Downloads

Court Mandates Executive Disclosure

Magistrate Judge Thomas Hixon has ordered Meta to disclose the names of executives who authorized the downloading of copyrighted books from shadow libraries like Anna’s Archive and LibGen. This ruling is part of discovery in class action lawsuits filed by authors, including Richard Kadrey and Sarah Silverman, and a separate lawsuit from publisher Entrepreneur Media.

Meta's Defense Strategy Challenged

Meta had previously attempted to avoid detailing its chain of command, particularly concerning leadership involved in early AI training dataset compilation. The company claimed an "informal approach" where no single authority made decisions, suggesting that staff "technically" downloaded torrents without explicit executive direction. This stance is now being challenged by the court order.

Evidence of Executive Approval

Testimony from Meta staff indicates that they were directed to download pirated content, with some expressing discomfort. One staffer noted that downloads "required some director or VP level approval to use for training." Another testified that leadership halted and then resumed downloading from Anna's Archive, contradicting Meta's claim of an informal, undirected process.

Implications for AI Training Practices

Authors and publishers argue that Meta made a deliberate choice to induce infringement, citing the massive scale of data downloaded (80.6 terabytes from LibGen and 35.7 terabytes from Anna’s Archive). They contend that such large-scale acquisition would necessitate seeding, which Meta claims to have mitigated. This case could set precedents for how companies acquire and use data for AI training, especially regarding copyrighted material.