🔍 Read the full analysis: What Investors Could Make Of Claims About Claude And Pirated Books on ThorstenMeyerAI.com
Get business pricing on office and shipping supplies
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
TL;DR
A 2025 ruling in the Bartz case found that training Claude on books was fair use, but that Anthropic’s downloading of more than seven million pirated book copies to build a library was not. Anthropic later agreed to a court-approved $1.5 billion settlement, while new lawsuits from authors who opted out remain unresolved. The company’s IPO filing discusses AI risks, but does not say Claude is sentient or will blackmail people.
Anthropic’s use of books to train Claude has produced a court ruling, a $1.5 billion settlement and new lawsuits, but those developments do not establish that every Claude model was trained on pirated books. In June 2025, a federal judge ruled that training on books was fair use while finding that downloading pirated copies to create a permanent library was not; the newer claims brought by authors who opted out of a settlement have not yet been decided.
The ruling came in Bartz v. Anthropic, a case filed by three authors in August 2024. In an order dated June 23, 2025, Judge William Alsup distinguished between using books to train a model and acquiring copies unlawfully. He found Anthropic’s model training to be fair use, but rejected the company’s downloading of pirated books to keep in a library. The order identified 196,640 books from Books3, at least five million copies from Library Genesis and at least two million from Pirate Library Mirror—more than seven million copies in total.
That decision was a pre-trial summary-judgment order, not a jury verdict on damages. The judge accepted the authors’ version of the facts about the pirated copies, while other questions were left for trial. The parties settled before a jury could decide them. The court approved a $1.5 billion settlement in July 2026, covering owners of roughly 500,000 registered books. The settlement provides about $3,000 per book before fees and requires Anthropic to destroy the downloaded collections and copies, subject to legal-hold obligations.
Anthropic has denied wrongdoing. In the settlement process, it represented that neither the LibGen nor PiLiMi datasets, nor portions of those datasets, were in the training corpus of any commercially released large language model. That statement and the earlier court record concern different questions: the record describes copying and training involving pirated collections, while the company’s settlement representation addresses commercially released models. The settlement releases specified past claims through August 25, 2025, but does not release claims about Claude’s outputs.
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
What the Case Means for Anthropic
For investors, the case is relevant because it connects copyright exposure to the cost and governance of building AI systems. The settlement puts a large, specific price on one group of claims and requires the destruction of the downloaded library. It does not, by itself, settle the legal status of all training practices or eliminate future claims involving other authors, datasets or model outputs.
The distinction between the court’s findings and Anthropic’s statements matters when judging risk. The judge found that acquiring the pirated copies was unlawful on the facts before him, but found the training itself fair use. Anthropic disputes wrongdoing and has said the identified datasets were not used in commercially released models. These positions should not be collapsed into either “a court ruled Claude was trained on pirated books” or “the court cleared every part of the process.”
The dispute also has potential bearing on how AI companies document their data sources and respond to rights holders. If courts or lawmakers impose additional requirements, companies could face new compliance costs, restrictions on training material or further litigation. The current record does not quantify those possible costs for Anthropic, and investors should treat them as unresolved risks rather than settled outcomes.
Popular right now
DeLonghi Rivelia Automatic Espresso Machine Bundle with Grinder and Frother, Beige
KitchenAid Artisan 5-Quart Tilt-Head Stand Mixer, Almond Cream
Dreame L40 Ultra Gen 2 Robot Vacuum and Mop
As an affiliate, we earn on qualifying purchases.
The headlines grew out of several separate developments. The Bartz case produced the June 2025 ruling, followed by a settlement covering registered owners of books in the specified pirated collections. Authors who did not join or accept that settlement have pursued separate litigation. Those filings are allegations, not findings that the new plaintiffs’ claims are valid.
A Stanford Daily report identified Stanford-affiliated writers whose books it said were among the pirated copies, including Tobias Wolff, Abraham Verghese, Chang-rae Lee and Daniel Mason. The source review behind this article says it did not independently verify the newspaper’s book list against the settlement database. It also notes a date and case distinction: the cited ruling was issued in June 2025 in Bartz, not in a May 2026 lawsuit. Chang-rae Lee is identified as a plaintiff in one new action, while the reviewer could not find Wolff among plaintiffs listed in five related cases as of August 2026.
Anthropic’s IPO filing, or S-1, is another distinct source. The source review says it discusses model behavior resembling blackmail in a staged safety test, the debate over possible AI sentience and existential concerns raised by others. It does not say Claude is sentient or that the company expects the system to blackmail people or end humanity. Those viral descriptions go beyond the filing as described in the source material.
““spectacularly so””
— Judge William Alsup, in the June 23, 2025 Bartz order
Questions the Court Has Not Resolved
The new lawsuits filed in May 2026 remain unresolved. The source material says Anthropic has moved to dismiss and a hearing is set for December 17, 2026; no ruling on those claims had been issued at the time described. The allegations in those cases should not be presented as court-established facts.
It is also unclear from the material provided how the plaintiffs’ claims will be affected by the settlement’s scope, what evidence the parties will present about particular books and models, or whether any additional cases will be filed. The source reviewer could not confirm the Stanford Daily’s book list against the settlement database and cautioned that the lists of plaintiffs in related suits may change through later amendments.
For investors, the S-1’s risk language is not a prediction that any described behavior will occur in ordinary use. The filing’s discussion of behavior resembling blackmail refers, according to the source review, to a staged safety test. The likelihood, business impact and regulatory consequences of the risks described in the filing are not quantified in the supplied material.
December Hearing and IPO Scrutiny
The next scheduled legal milestone is the December 17, 2026 hearing on Anthropic’s motion to dismiss the new author claims. A hearing is not a ruling; the court may issue a decision later, and the source material does not specify when that might happen. Any decision could clarify which claims proceed, but it would not automatically resolve broader copyright questions involving other cases or datasets.
Investors can also watch for developments in Anthropic’s public-offering process and how the company describes litigation, data sourcing and AI-related risks in its filings. The settlement payment schedule extends through September 2027, according to the source material, while the company is required to destroy the specified downloaded books subject to legal-hold obligations. Further court filings, amendments to the new cases or updated disclosures could change what is known.
For now, the clearest reading is narrow: the 2025 court order treated training and pirated acquisition differently; the settlement resolved specified past claims without deciding every issue; and the new author cases remain pending. No material supplied here establishes that Claude is sentient or that it will blackmail users.
Key Questions
Did a judge rule that Claude was trained illegally on books?
Not in those terms. In June 2025, Judge William Alsup found that training on books was fair use, but ruled that downloading pirated copies to build a permanent library was not. The order addressed the facts in the Bartz case and did not decide every question about every Claude model.
What does Anthropic’s $1.5 billion settlement cover?
The July 2026 court-approved settlement covers owners of roughly 500,000 registered books that Anthropic downloaded from LibGen or PiLiMi, according to the source material. It provides about $3,000 per book before fees, requires destruction of the downloaded books and copies subject to legal holds, and releases specified past claims—not claims about Claude’s outputs.
Are the new author lawsuits against Anthropic decided?
No. The lawsuits described in the source material were filed in May 2026, and the court had not ruled on their claims. A hearing on Anthropic’s motion to dismiss is scheduled for December 17, 2026.
Does Anthropic’s IPO filing say Claude is sentient or will blackmail people?
No, according to the source review. The filing discusses behavior resembling blackmail in a staged safety test, debate over possible AI sentience and existential concerns raised by others. That is not a statement that Claude is sentient or will blackmail people.
Source: ThorstenMeyerAI.com
