🔍 Read the full analysis: Claude’s Training Data Under Scrutiny: The Anthropic Book Allegations on ThorstenMeyerAI.com
Get the latest gadgets delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
TL;DR
A 2025 federal court ruling found that training Claude on books was fair use, while separately finding that Anthropic’s downloading and storage of pirated book copies was not. Anthropic later agreed to a $1.5 billion settlement over the piracy claims; separate lawsuits filed by authors who opted out have not been decided. The company’s S-1 filing discusses model behavior and AI-related risks, but does not say Claude is sentient or will blackmail people.
Fresh scrutiny of Anthropic’s use of books to train Claude is renewing debate over a distinction at the center of its copyright cases: a federal judge found in 2025 that training on books was fair use, but ruled that downloading and keeping pirated copies was not. Anthropic later agreed to a $1.5 billion settlement covering the piracy claims, while separate lawsuits by authors who opted out remain unresolved.
In Bartz v. Anthropic, U.S. District Judge William Alsup issued a summary-judgment order on June 23, 2025. He found that using books to train Claude was fair use and also treated Anthropic’s purchase and scanning of used print books as fair use. Separately, he found that obtaining pirated copies to build a permanent digital library was not fair use. The order described more than seven million book copies from the Books3, Library Genesis and Pirate Library Mirror collections.
The piracy finding was not a final jury verdict on damages. The judge accepted the authors’ version of the facts for that part of the pretrial decision, and a jury was expected to address remaining issues. The case settled before trial. The settlement, approved in July 2026 according to the source material, provides $1.5 billion for owners of roughly 500,000 registered books, or about $3,000 per book before fees. It also requires Anthropic to destroy the downloaded LibGen and PiLiMi books and copies, subject to legal-hold obligations.
Anthropic denies wrongdoing. In settlement materials, the company represented that neither the LibGen nor PiLiMi datasets, nor portions of them, were in the training corpus of any commercially released large language model. The settlement addresses past conduct through August 25, 2025, and does not release claims about Claude’s outputs. Authors who opted out have filed separate suits; the source says a hearing on Anthropic’s motion to dismiss is set for December 17, 2026.
From download to IPO
Five years, one ruling, no trial
The piracy happened in 2021 and 2022. The only ruling came in 2025. The new lawsuits are years away from a verdict.
- Jun 2021 · Jul 2022DownloadsAt least 5 million copies from LibGen (2021), at least 2 million from PiLiMi (2022); 196,640 from Books3 earlier.
- Aug 2024Bartz v. Anthropic filedThree authors sue Anthropic, on behalf of a class of book owners.
- 23 Jun 2025Judge Alsup's orderTraining on books: fair use. Scanning bought books: fair use. Pirated library copies: not fair use, to trial.
- Sep 2025$1.5 billion settlementNo trial. Anthropic denies wrongdoing; release covers conduct up to 25 Aug 2025 only.
- 13 + 15 May 2026Opt-out suitsCruz (28 plaintiffs) and Kwon (incl. Chang-rae Lee) join Cambronne and two publisher suits.
- Jul 2026Final approvalSettlement approved. Anthropic: no opt-out deal "above the Bartz claim value".
- 24 Sep 2026Motion to dismissAnthropic: claims too late, no infringing output shown, many works already released.
- 26–27 Oct 2026NowS-1 and the headlineIPO filing cites Bartz as a risk; The Stanford Daily names Stanford authors.
- 17 Dec 2026ScheduledHearingCourt hears the motion to dismiss.
- May 2027 · Jan 2028ScheduledDiscovery, then trial settingFact discovery closes 13 May 2027; trial-setting conference 11 Jan 2028.
Sources: Bartz v. Anthropic, order of 23 June 2025; settlement FAQ; joint case statement, 27 Aug 2026; Anthropic letter, 27 Jul 2026; case management order, 14 Sep 2026; Anthropic S-1, 26 Oct 2026; The Stanford Daily, 27 Oct 2026. Scheduled dates can move.
Powered by Thorsten Meyer AI
The Difference Between Training and Piracy
The dispute matters because the court’s conclusions do not amount to a blanket ruling that all AI training on copyrighted books is lawful, nor do they establish that every book in a pirated collection was used to train a released model. The judge drew a distinction between learning from books through training and acquiring and retaining pirated copies. That distinction is central to how authors, AI developers and other courts may assess copyright disputes.
The settlement also has practical consequences: it provides compensation to eligible book-rights holders and requires destruction of the specified collections. But it does not settle every dispute about Anthropic’s conduct. The opt-out cases can still test claims outside the settlement, while the company’s stated position on the training corpora remains an important part of the record. For readers, the careful takeaway is that the court found piracy-related conduct unlawful, but separately found the training at issue fair use.
Products Worth Considering

Samsung 27-Inch Odyssey G5 QHD Gaming Monitor (G50F)

LG 27GR83Q-B 27-inch Ultragear QHD IPS Gaming Monitor

Apple MacBook Air 13-inch with M5 Chip, 16GB Memory, 512GB SSD, Midnight
As an affiliate, we earn on qualifying purchases.
How the Cases Reached This Point
Bartz v. Anthropic was filed by three authors in August 2024. In June 2025, Judge Alsup’s order separated Anthropic’s book-scanning and model-training practices from its acquisition of pirated digital copies. He described the training as transformative, while finding that the copying and permanent storage of pirated books could not be excused by buying copies later.
The authors’ settlement followed, setting compensation for registered books in the covered collections and resolving specified past claims. A separate group of authors who did not join the settlement has pursued new litigation. The source material reports that novelist and Stanford professor emeritus Tobias Wolff was among the Stanford authors identified by The Stanford Daily as having works in the pirated copies, but says his name was not found among plaintiffs listed in five related cases as of August 2026. The book list has not been independently checked against the settlement database in the source account.
There is also a separate discussion around Anthropic’s initial public offering filing. The source says the S-1 describes model behavior that “resembling blackmail” in a staged safety test, possible regulatory debate over AI sentience, and concerns others have raised about existential risks. Those passages are not statements that Claude is sentient or will blackmail users.
““neither the LibGen or PiLiMi datasets, nor any portions of those datasets, were in the training corpus of any of its commercially released large language models.””
— Anthropic, as described in the settlement materials
Claims Still Awaiting Court Review
The opt-out authors’ claims have not yet been decided. The scheduled December 17, 2026 hearing concerns Anthropic’s motion to dismiss; its outcome and any later trial schedule are not established in the supplied material. It is also not clear from that material which authors’ works will be verified as eligible for payment, or whether Wolff is part of another action or a later amendment.
The record described here also leaves a distinction readers should keep in view: the 2025 order addressed the copying and use of pirated collections in the case before it, while Anthropic’s settlement representation says those datasets were not in commercially released models’ training corpora. The supplied account does not resolve every question about which specific books entered which model-development stages. Nor does it establish a court finding that Claude’s outputs infringed authors’ rights; the settlement explicitly leaves output-related claims unreleased.
The December Hearing and Settlement
The next stated court milestone is the December 17, 2026 hearing on Anthropic’s motion to dismiss the opt-out authors’ cases. The court’s decision will show whether those claims proceed, though the supplied information does not predict the result. The settlement also calls for a final $450 million instalment by September 27, 2027, following earlier payments specified in its terms.
Readers should watch for court orders, updated plaintiff lists and settlement-administration information rather than treating new headlines as new rulings. Any reporting on the S-1 should likewise distinguish the filing’s risk language from claims that the company or a court has confirmed Claude is sentient or will blackmail people.
Key Questions
Did a judge rule that all of Anthropic’s book training was illegal?
No. Judge William Alsup found that training Claude on books was fair use in the 2025 case. He separately found that downloading and keeping pirated copies was not fair use.
What does the $1.5 billion settlement cover?
It covers specified past claims tied to pirated book copies from LibGen and PiLiMi, with payments for owners of roughly 500,000 registered books. The settlement also requires destruction of the downloaded collections and copies, subject to legal-hold obligations.
Have the opt-out authors’ lawsuits been decided?
No. The cases remain pending. The supplied source says a hearing on Anthropic’s motion to dismiss is scheduled for December 17, 2026.
Does Anthropic’s S-1 say Claude is sentient or will blackmail people?
No, according to the source material. It says the filing discusses behavior resembling blackmail in a staged safety test, possible regulation related to debate over AI sentience, and existential concerns raised by others. Those are not claims that Claude is sentient or will blackmail users.
Source: ThorstenMeyerAI.com
