🔍 Read the full analysis: How To Separate Claims From Evidence In The Anthropic Book Dispute on ThorstenMeyerAI.com
Get business pricing on monitors, keyboards and dev gear
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
TL;DR
A federal judge ruled in June 2025 that Anthropic’s use of books to train Claude was fair use, while finding that the company’s downloading and retention of pirated book copies was not. A $1.5 billion settlement over those copies was approved in July 2026; separate lawsuits by authors who opted out have not been decided.
Anthropic’s book dispute has produced a court finding of unlawful copying, a separate ruling that training Claude on books was fair use, and a $1.5 billion settlement covering pirated copies. Those findings do not amount to a ruling that all Claude training was illegal: new lawsuits from authors who did not join the settlement remain unresolved, with a hearing scheduled for December 17, 2026.
In Bartz v. Anthropic, U.S. District Judge William Alsup ruled on June 23, 2025, that training Claude on books was fair use under copyright law. He also found that Anthropic’s scanning of print books it had purchased was fair use. But the ruling treated a different practice separately: downloading pirated copies and keeping them in a permanent library. Alsup found that conduct was not fair use. The order described more than seven million copies from collections including Books3, Library Genesis and the Pirate Library Mirror.
The decision was a pretrial ruling, not a jury verdict on damages. On the pirated copies, Alsup accepted the authors’ account of the facts; the case did not proceed to a jury trial because Anthropic settled. The settlement, approved in July 2026, 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 requirements.
The settlement covers specified past conduct through August 25, 2025; it does not release claims about Claude’s outputs. Anthropic denies wrongdoing. In the settlement, the company represented that the LibGen and PiLiMi datasets, or portions of them, were not in the training corpus of any commercially released model. That representation should be distinguished from the earlier ruling, which found that pirated copies had been downloaded and that portions of the collections were used to train some Anthropic models. The settlement was not a trial verdict resolving every question about training data.
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 Settlement Resolves
The distinction matters because copying books, training on books and distributing model outputs are different legal questions. The 2025 order addressed training and the acquisition and retention of copies on the record before the court. The settlement resolved covered claims about specified past conduct without licensing every book or deciding future disputes. It also does not establish that every author or book cited in news coverage was part of the settlement class.
For authors and publishers, the settlement sets a substantial payment for registered books tied to the covered pirated collections, while leaving other claims open. For AI companies, the dispute shows that a court can regard model training as fair use while finding the acquisition or storage of training materials unlawful. Those conclusions are specific to this case and do not settle how courts will assess other datasets or models.
Readers also have reason to separate the book litigation from claims about model behavior. Anthropic’s public-market filing describes models showing behavior “resembling blackmail” in testing and refers to public concerns about AI sentience and existential risks. The filing does not state that Claude is sentient or that it will blackmail people. Describing a risk disclosure as a prediction or confirmed outcome would go beyond what the document says.
Products Worth Considering
Apple iPad Air (2025) 13‑inch Wi‑Fi 256 GB – Space Gray (M3 chip)
LG 34GX90SA-W 34-inch Ultragear WQHD OLED Curved Gaming Monitor
Bose S1 Pro+ Portable PA Speaker
As an affiliate, we earn on qualifying purchases.
How the Cases Fit Together
The original case, Bartz v. Anthropic, was filed by three authors in August 2024. In his June 2025 order, Alsup distinguished between books Anthropic bought and scanned, books used to train Claude, and pirated files kept in a library. His finding that training was fair use did not excuse what he characterized as downloading pirated books to build a permanent collection. The settlement later addressed claims over the covered books and specified past conduct.
Authors who opted out have brought separate lawsuits. According to the source material, a related case names author Chang-rae Lee as a plaintiff. The Stanford Daily reported that works by several Stanford-associated authors, including Tobias Wolff, were among the affected books, and that ten Wolff works were involved. Those details are the newspaper’s reporting; the underlying book list has not been independently verified here. The source review also says it could not find Wolff among plaintiffs listed in five related cases as of August 2026, so his participation in a separate action or later filing is not established by that review.
Some coverage has described a judge ruling in a May 2026 lawsuit that Anthropic pirated the books. The court order cited in the reporting is instead dated June 2025 and belongs to the earlier Bartz case. The newer author claims should not be presented as though they have already received a judicial ruling.
““spectacularly so””
— Judge William Alsup, in the June 23, 2025 order
Which Claims Still Lack Rulings
The opt-out authors’ cases are not yet decided. A hearing on Anthropic’s motion to dismiss is scheduled for December 17, 2026; the hearing date does not itself indicate how the judge will rule. The source material does not establish whether the court will dismiss any claims, allow them to proceed, or when a later trial might occur.
There is also a gap between the settlement and broad claims about particular authors’ books. The reported list of affected works has not been checked against the settlement database in the source account, and the available account does not confirm every author’s status in the related lawsuits. Nor does the settlement representation establish, by itself, what data was used to train every model or answer all questions about the history of the datasets. Those points require care rather than assumptions based on headlines.
Finally, the court’s fair-use ruling is tied to the facts and legal arguments in Bartz. It should not be read as a blanket decision that any method of collecting books, any training use, or any model output is lawful. Claims about the prospectus and AI risks also need to be checked against its wording: risk disclosures describe possibilities and concerns, not necessarily events that have happened.
The December Hearing and Beyond
The next scheduled milestone identified in the source material is the December 17, 2026 hearing on Anthropic’s motion to dismiss the opt-out authors’ claims. The court’s action could clarify which claims may continue, but the outcome and any later schedule remain unknown. Any further assessment should distinguish the new cases from the claims resolved by the Bartz settlement.
Readers following the dispute can check future reporting against primary documents: the dated court orders, the final settlement terms, the complaints and motions in the newer cases, and Anthropic’s own filings. The clearest account at present is limited: the judge ruled that training was fair use, found the covered pirated downloads unlawful on the record before him, and the settlement resolved specified claims. The remaining lawsuits and broader questions about other books, datasets and models still require separate evidence and court decisions.
Key Questions
Did a judge rule that Claude’s book training was illegal?
No. In June 2025, Judge William Alsup ruled that training Claude on books was fair use. He separately found that downloading and keeping pirated book copies was not fair use.
What does the $1.5 billion settlement cover?
The settlement covers specified claims involving roughly 500,000 registered books downloaded from LibGen or PiLiMi and past conduct through August 25, 2025. It does not release claims about Claude’s outputs or grant a general license to train on books.
Have the new author lawsuits been decided?
No ruling on the opt-out authors’ new claims is reported in the source material. A hearing on Anthropic’s motion to dismiss is scheduled for December 17, 2026.
Does Anthropic’s filing say Claude is sentient or will blackmail people?
No. The filing refers to model behavior “resembling blackmail” and discusses debate about sentience and existential risks. Those are risk disclosures and descriptions of concerns, not a statement that Claude is sentient or will blackmail people.
Source: ThorstenMeyerAI.com
