Meta Argues BitTorrent Seeding of Pirated Books for LLM Training Is Fair Use

Meta has filed a supplemental interrogatory response in the ongoing Kadrey v. Meta copyright lawsuit arguing that its automatic uploading of pirated books to other BitTorrent peers during LLM training data acquisition qualifies as fair use, because seeding is an inseparable technical consequence of the BitTorrent download protocol. The company previously won the core training fair use claim in summer 2025; the BitTorrent distribution claim is the only remaining live issue in a case filed in 2023. Authors' counsel filed an emergency letter arguing Meta's fair use upload defense was improperly raised after the discovery deadline, which Meta disputes by citing a December 2025 joint case management statement.

Key Takeaways

• Meta won training fair use in summer 2025; remaining live issue is direct copyright infringement from BitTorrent seeding — Meta now argues seeding was "part-and-parcel" of the transformative training download and therefore also fair use
• Meta's own deposition evidence shows named authors including Sarah Silverman admitted they are unaware of any Meta model output replicating their books, and that it "doesn't matter at all" if models never output their content
• Case before Judge Vince Chhabria in the Northern District of California; if Meta's "fair use by technical necessity" defense is upheld, it would significantly affect the many other AI copyright lawsuits involving shadow library torrent data

Original source: TorrentFreak