Two lawsuits named a shadow library as an AI training source
Silverman, Kadrey and Golden sued OpenAI and Meta the same day, both complaints citing the Books3 dataset.
- Historical event
- July 7, 2023
- First source published
- July 7, 2023
- Site publication
- September 18, 2026
The record
On 7 July 2023, authors Richard Kadrey, Sarah Silverman and Christopher Golden filed two proposed class-action complaints in the Northern District of California, one against Meta Platforms and one against OpenAI, on behalf of themselves and a proposed class of writers. The complaint against Meta alleges direct and vicarious copyright infringement arising from Meta's use of a training dataset called Books3 to train its LLaMA language models; the complaint against OpenAI, filed the same day, makes parallel allegations concerning OpenAI's language models.
What the documents show
Both filed complaints identify Books3 by name, tracing it to a paper published by the research organisation EleutherAI describing a broader dataset called "The Pile." The Meta complaint quotes that paper's own description of Books3 as "a dataset of books derived from a copy of the contents of the Bibliotik private tracker," and both complaints describe Bibliotik as one of several "shadow library" sites, alongside Library Genesis, Z-Library and Sci-Hub, that the filings characterise as unauthorised sources of copyrighted books later used in bulk. The complaints attach exhibits listing plaintiffs' books the filings say appear within Books3. These are allegations set out in a filed complaint, not findings by any court; whether the described use of Books3 constitutes infringement, and what if anything a defendant admits or disputes about it, is a separate question the complaints themselves do not resolve.
Why it stays interesting
The two complaints matter to a publishing reader independent of their eventual outcome, because they put a specific, named, and independently documented dataset at the center of an AI-training dispute rather than a general allegation about "AI companies." A reader assessing any later reporting on AI training-data litigation can use these complaints as a baseline: what dataset is actually named, what document traces its contents, and what the filing itself, rather than commentary about it, claims.
What to read next
The Authors Guild's own model contract clause, drafted the same year to let authors withhold their books from AI training by contract, addresses the same underlying concern through a different mechanism, a negotiated term rather than a lawsuit, and is worth reading against these complaints for contrast.
- What does the cited EleutherAI paper itself say about how the Books3 dataset was assembled, separate from what either complaint alleges about it?
- How does a claim of direct infringement differ, on the face of a complaint, from a claim of vicarious infringement, and what must each eventually show?
- What subsequent filings, rulings or settlements in this docket would confirm, narrow or dispose of the claims described here?
Filed the same day against two different companies, the complaints share a dataset, a set of plaintiffs and a legal theory, and reading them together shows how one contested dataset became the pleaded center of parallel cases.
Sources & reading trail
The filed complaint's allegations, its plaintiffs, and its identification of the Books3 dataset and Bibliotik shadow library.
Source published: 7 July 2023 · Retrieved: 16 September 2026
Confirms the companion complaint filed the same day against OpenAI, naming the same plaintiffs and dataset.
Source published: 7 July 2023 · Retrieved: 16 September 2026
Reports, filings, records and studies establish the entry; the reading of why it stays interesting is Chapter Ledger editorial analysis. This retrospective draft does not imply the site published on the event date.
Continue reading
- Seventeen novelists sued an AI company as a proposed class
- A federal court found library ebook scanning was not fair use
- Authors Guild drafted a clause to block AI training uses
- Browse the complete the ledger archive
Sources & reading trail
- Kadrey et al. v. Meta Platforms, Inc., Class Action Complaint (N.D. Cal. No. 3:23-cv-03417)
Source published: July 7, 2023 · Retrieved: September 16, 2026 - Silverman et al. v. OpenAI, Inc., Class Action Complaint (N.D. Cal. No. 3:23-cv-03416)
Source published: July 7, 2023 · Retrieved: September 16, 2026
The documents above establish the record. The reading and the questions are this publication’s editorial analysis, written after the fact.
Published September 18, 2026, not on the date of the event described.