Meta Ordered to Disclose Staff Behind Book Torrenting for AI Training

Meta must identify the internal staff members who authorized downloading copyrighted book repositories over BitTorrent networks, according to reporting from TorrentFreak on a recent ruling by Magistrate Judge Thomas Hixson.

The discovery order comes amid ongoing copyright litigation focused on how the tech giant sourced training data for its initial Llama artificial intelligence models. While the court previously determined that training neural networks on copyrighted material constitutes fair use, Meta remains entangled in contributory infringement claims concerning whether its servers seeded or distributed pirated files to third parties during downloads from shadow libraries such as LibGen and Anna's Archive.

Entrepreneur Media, the publishing company behind Entrepreneur magazine, initiated the push to unmask the decision-makers behind the data pipeline. The publisher demanded that Meta name every employee who made, authorized, or approved obtaining shadow library datasets through BitTorrent protocols.

Meta had attempted to avoid singling out individual decision-makers, asserting that choices regarding data sourcing for its first two Llama releases resulted from broad team consensus rather than formal corporate authorizations, TorrentFreak reported. Judge Hixson had already instructed the company to surrender server command histories, which are expected to reveal how Meta configured its torrent clients and whether upload mechanisms were enabled.

As Hollywood Tomorrow tracks the evolving intellectual property fights surrounding generative technology, the ruling shows how data acquisition logistics, rather than pure model training, represent an immediate legal exposure for tech firms. Pinning individual responsibility on engineers and product managers could significantly complicate tech industry defenses that treat systemic scraping as routine development work.