Kadrey v. Meta
The LLaMA Copyright Lawsuit
Meta trained its LLaMA AI models on Books3 — a dataset of ~196,000 pirated books from shadow libraries — without permission or payment. Richard Kadrey, Sarah Silverman, and the Authors Guild are among the plaintiffs.
Case Overview
Meta's LLaMA family of large language models — including LLaMA, LLaMA 2, LLaMA 3, and Llama 3.1 — were trained on datasets that included Books3, a collection of approximately 196,000 full-text books sourced from shadow libraries. Books3 was assembled without licensing from copyright holders and consists largely of pirated ebook files. It was also part of LibGen (Library Genesis), the notorious piracy site that was also central to the Anthropic litigation.
In 2023, authors Richard Kadrey (known for the Sandman Slim urban fantasy series), Christopher Golden (horror author), and Sarah Silverman (comedian and memoirist) filed the initial complaint against Meta in the Northern District of California. The complaint alleged that Meta copied and used their copyrighted works without permission as training data for LLaMA.
The Authors Guild and other organizations subsequently filed related actions. The consolidated litigation now represents a broad class of authors whose books appeared in Meta's training datasets. The case centers on whether using pirated book datasets for AI training constitutes copyright infringement — the same legal question at the heart of the Anthropic settlement.
Current Status
What We Know
How This Compares to Anthropic
Similarities
- ✓ Same datasets (Books3, LibGen)
- ✓ Same legal theory (copyright infringement via pirated training data)
- ✓ Same court (N.D. California)
- ✓ Similar class size potential (~100K+ eligible works)
Key Differences
- → Meta has not settled — litigation is ongoing
- → Meta is a larger company with deeper pockets
- → LLaMA is open-source — different fair use arguments
- → Timeline to settlement unknown
The Anthropic settlement ($1.5B for ~400K books) is a useful reference point. If Meta settles on similar terms, qualifying authors could see comparable per-work payouts.
Be First to Know
Join our waitlist and we'll notify you the moment a Meta/LLaMA settlement is announced or claims become available. No spam — just updates that matter.
FAQ
What is Books3 and why does it matter?
Books3 is a dataset of approximately 196,000 full-text books assembled from shadow libraries — primarily Bibliotik, a private torrent tracker for ebooks. It was compiled by researcher Shawn Presser and published as part of The Pile dataset. Meta used Books3 (among other datasets) to train its LLaMA models. Because Books3 consists of pirated books, its use in AI training raises significant copyright infringement questions.
Who filed the Kadrey lawsuit?
Richard Kadrey (author of the Sandman Slim series), Christopher Golden (horror and fantasy author), and Sarah Silverman (comedian and author) filed the initial complaint against Meta. The Authors Guild and other organizations have since filed related actions. The case is proceeding in the Northern District of California.
What is the current status of Kadrey v. Meta?
As of April 2026, the case is in active litigation — no settlement has been reached. The case has advanced through early procedural stages and is in the discovery phase. Meta has contested several claims, but the core copyright infringement allegations remain active.
How does this compare to the Anthropic case?
The fundamental allegation is nearly identical: an AI company downloaded pirated books (primarily from Books3/LibGen) and used them to train AI models without licensing or compensating authors. The scope is similar — Books3 contains ~196,000 books. If Meta reaches a settlement, the structure would likely resemble Anthropic's: a large fund distributed pro rata to qualifying copyright holders.
What should I do right now?
Join our waitlist. We'll notify you immediately when a settlement is announced or when claims become available. In the meantime, check if your books qualify for the Anthropic settlement — which is already approved and moving toward distribution.
Have books that may qualify for the Anthropic settlement?
That settlement is already approved and moving toward distribution. Check if your books are on the Anthropic Works List — same datasets, same authors, paying out now.
Explore Bartz v. Anthropic →