Newly released court documents show Meta downloaded at least 81.7 terabytes of stolen books through torrenting to develop its AI models, which has triggered major copyright issues. Book authors submitted internal Meta communications that demonstrate that the company deliberately participated in massive unauthorized data copying.
Meta had admitted to using LibGen as a massive pirated book collection, but the precise scope of its activities remained unknown until these recent court documents were unsealed. Meta obtained its pirated book data through direct downloads from both Z-Library and Anna’s Archive as well as other shadow libraries. The court documents demonstrate that Meta used LibGen to download 80.6TB of files in an independent incident.
The legal situation for Meta grew more serious when internal emails exposed that company workers understood the potential risks. The April 2023 message from Meta research engineer Nikolay Bashlykov showed his concern about corporate laptop torrenting while he added a smiley emoji for humor. In September he approached the legal team because he believed that seeding the files might violate copyright regulations. Meta continued its torrenting activities instead of stopping them by routing its network through different servers while reducing file sharing participation.
New evidence shows that Mark Zuckerberg and other Meta executives participated in choosing LibGen as per previous author testimony. Meta continues to assert that LibGen training falls within “fair use” boundaries, yet the company faces growing legal scrutiny.
Meta denies distributing any copyrighted books, but authors maintain that their seeding activities support their copyright infringement claim. Discovery activities have forced Meta to defend itself against accusations of actively concealing its torrenting operations as the company profited from stolen content.
