🔍 Read the full analysis: Did Anthropic Train Claude On Copyrighted Books? What’s Alleged 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 federal judge ruled in June 2025 that Anthropic’s use of books to train Claude was fair use, while finding that the company’s retention of millions of pirated book copies was not. Anthropic later agreed to a $1.5 billion settlement covering about 500,000 books, and new lawsuits filed by authors in 2026 have not been decided.
Anthropic’s use of books to train Claude was ruled fair use in a June 2025 federal court order, but the judge found that the company’s downloading and permanent storage of millions of pirated books was not excused by fair use. That distinction is central to new lawsuits filed by authors in 2026, whose allegations remain unresolved; Anthropic also agreed to a $1.5 billion settlement over roughly 500,000 books.
In Bartz v. Anthropic, a case brought by three authors in 2024, U.S. District Judge William Alsup issued a summary-judgment order on June 23, 2025. He concluded that training Claude on books could qualify as fair use and described the use as “spectacularly” fair. He also found that scanning books Anthropic had bought was fair use. Those findings did not cover every way the company obtained or retained books.
The court separately addressed Anthropic’s digital library. Its order described 196,640 books from Books3, at least five million copies from Library Genesis (LibGen), and at least two million from the Pirate Library Mirror (PiLiMi). Alsup found that downloading and keeping pirated copies was not fair use. He accepted the authors’ account of the relevant facts for that pretrial decision; a jury was expected to address remaining issues, including damages, before the parties settled.
The settlement received final approval in July 2026 and provides $1.5 billion for owners of about 500,000 registered books, or roughly $3,000 per book before fees. Anthropic disputes wrongdoing. In the settlement, it 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 agreement also requires destruction of the downloaded collections and copies, subject to legal-hold obligations. It resolves specified past conduct through August 25, 2025, but does not release claims concerning 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
The Difference Between Training and Piracy
The distinction affects both the legal claims and how readers should interpret the headline that Anthropic “illegally trained” Claude on books. The 2025 ruling did not find that the act of training on books was unlawful; it found that retaining pirated copies in a permanent library was not fair use. Saying the judge ruled that Claude’s training was illegal collapses two separate issues into one.
The settlement has a large financial consequence for authors whose books were registered, but it is not a general licence for AI companies to use copyrighted works. Nor does it settle every dispute about how Claude was trained or what it produces. Authors, publishers and AI developers are watching how courts distinguish access to copyrighted material, its use in training, and the storage of unauthorized copies.
Products Worth Considering
DJI Osmo Mobile 8 3-Axis Smartphone Gimbal Stabilizer
Logitech MX Mechanical Wireless Illuminated Keyboard, Tactile Quiet, Graphite
Logitech MX Keys S Wireless Keyboard, Graphite
As an affiliate, we earn on qualifying purchases.
From the 2025 Order to New Suits
The original Bartz case began in August 2024. In his June 2025 order, Alsup drew a line between Anthropic’s purchase and scanning of physical books, which he found fair use, and its acquisition of pirated digital collections. He wrote that later buying a copy would not erase liability for the earlier theft. Because the case settled, the anticipated jury trial on remaining questions did not take place.
Authors filed new suits in May 2026 after opting out of the settlement, according to the source material. One case, Kwon et al. v. Anthropic, lists author Chang-rae Lee among its plaintiffs. The Stanford Daily also reported that books by Stanford-affiliated writers, including Tobias Wolff and Abraham Verghese, were among the pirated copies. That reported list has not been independently verified against the settlement database in the material provided, and Wolff’s name was not found among plaintiffs in the five related cases reviewed as of August 2026.
A further source of confusion is timing: the 2025 court order and 2026 suits are separate developments. No judge issued a new ruling in May 2026 on the authors’ claims. Anthropic’s public offering filing is also a separate document; claims circulating online about it should be checked against its actual wording rather than presented as court findings.
““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, in its settlement representation
Claims Awaiting a Court Ruling
The new authors’ cases are allegations, not findings of liability. The source material says Anthropic denies wrongdoing and that a hearing on its motion to dismiss is scheduled for December 17, 2026. The court has not yet ruled on those claims, and the eventual scope of any case that proceeds is not known.
The available record also leaves unresolved which specific titles were included in particular collections and which were used in training particular models. The settlement covers registered books in the specified pirated collections, not every copyrighted book or every possible use of a book. Anthropic’s statement about commercially released models should be attributed to the company; it does not by itself resolve disputes over other models, training stages, or the new plaintiffs’ claims.
Reporting about Stanford-affiliated authors and their books should likewise be treated carefully. The source material says the list was reported by The Stanford Daily but not checked against the settlement database, and it could not confirm Wolff as a named plaintiff in the five related cases reviewed as of August 2026.
December Hearing and Case Rulings
The next stated court milestone is the December 17, 2026 hearing on Anthropic’s motion to dismiss the new authors’ case. The hearing could determine whether some or all of the claims proceed, but it will not itself establish that Anthropic infringed copyright. Any decision should be distinguished from Alsup’s 2025 order and the settlement in Bartz.
Anthropic’s settlement payments are scheduled in instalments, with a final stated payment of $450 million due by September 27, 2027. The agreement also requires destruction of the covered pirated collections, subject to legal-hold requirements. Further court filings or rulings may clarify which authors and works are involved and what claims remain in dispute.
Key Questions
Did a judge rule that Anthropic illegally trained Claude on books?
No. In June 2025, Judge William Alsup found that training Claude on books was fair use. He separately ruled that downloading and keeping pirated copies was not fair use.
What did Anthropic agree to pay?
Anthropic agreed to a $1.5 billion settlement for owners of roughly 500,000 registered books from the specified LibGen and PiLiMi collections. The settlement received final approval in July 2026.
Have the authors’ 2026 lawsuits been decided?
No. The claims filed in May 2026 remain unresolved. A hearing on Anthropic’s motion to dismiss is scheduled for December 17, 2026, according to the source material.
Does the settlement resolve claims about Claude’s answers?
No. The settlement addresses specified past conduct through August 25, 2025, and does not release claims concerning Claude’s outputs.
Did Anthropic say pirated books were used to train released models?
Anthropic represented in the settlement that the LibGen and PiLiMi datasets, or portions of them, were not in the training corpus of any commercially released large language model. That is the company’s statement; the new lawsuits have not been adjudicated.
Source: ThorstenMeyerAI.com
