🔍 Read the full analysis: Anthropic’s Book Piracy Claims: Evidence, Uncertainty, And IPO Concerns 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 court order found Anthropic’s use of books to train Claude was fair use, but said the company’s downloading and retention of more than seven million pirated book copies was not. Anthropic later agreed to a $1.5 billion settlement, while new lawsuits by 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 has filed to go public, bringing renewed scrutiny to its book-piracy settlement and to new copyright lawsuits that have not been decided. A June 2025 court order found that training Claude on books was fair use, but ruled that Anthropic’s downloading and retention of more than seven million pirated book copies was not; the company later agreed to a $1.5 billion settlement.
The ruling came in Bartz v. Anthropic, a case brought by three authors in August 2024. In an order dated June 23, 2025, U.S. District Judge William Alsup distinguished between how the books were used. He found that training Claude on books, and scanning books Anthropic had bought, qualified as fair use. But the order said keeping pirated copies in a permanent library did not. The court record described 196,640 Books3 titles, at least five million copies from Library Genesis and at least two million from the Pirate Library Mirror.
The order was a pretrial decision, not a jury verdict on damages. On the pirated-copy issue, Alsup accepted the authors’ version of the facts; a jury was expected to decide remaining matters, including damages. The case settled before trial. The settlement, finally approved in July 2026, provides about $3,000 for each of roughly 500,000 registered books, 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, the company represented that the LibGen and PiLiMi datasets, or portions of them, were not in the training corpus of any commercially released large language models. The new suits are separate: authors who opted out have brought further claims, and no ruling on those claims had been issued as of the source report. A hearing on Anthropic’s motion to dismiss is scheduled 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
How the Settlement Shapes Anthropic’s IPO
The distinction between training and acquisition matters to publishers, authors and AI companies. The 2025 order did not find that training Claude on books was copyright infringement; it found that obtaining and retaining pirated copies was not fair use. Treating the ruling as a blanket finding that “AI training was illegal” would misstate what the court decided.
The settlement is also a significant financial and legal outcome for Anthropic. It provides payments to eligible rights holders, but it does not settle every dispute about the company’s use of copyrighted material. The agreement releases claims tied to past conduct only through August 25, 2025; claims about Claude’s outputs are expressly not released. The opt-out cases can still test questions not resolved by the settlement.
For prospective investors, the timeline and scope are relevant disclosures, not proof of how the remaining cases will end. The company’s IPO filing also describes potential AI safety and regulatory risks. Viral accounts that say the filing admits Claude could become sentient or will blackmail people go beyond the wording described in the source report.
Popular right now
Samsung 42-Inch Class OLED S90F 4K Smart TV (2025)
Dyson V8 Cyclone Cordless Vacuum
Shark Navigator Lift-Away Upright Vacuum (NV360)
As an affiliate, we earn on qualifying purchases.
What the 2025 Court Order Decided
Judge Alsup’s order drew a line between copying for model development and the source of the copies. He compared training to learning from reading and described Claude’s training as an effort “to turn a hard corner and create something different.” He separately found Anthropic’s scanning of purchased books fair use. By contrast, the court said downloading pirated copies to keep was not fair use, and that buying a copy later would not erase liability for the earlier copying.
The $1.5 billion settlement covers registered books from LibGen and PiLiMi and is to be paid in instalments. Its release is limited: it addresses past conduct through the specified cutoff date, not every copyright question involving Claude. Anthropic’s representation about the training corpus is part of the settlement record; it should be read alongside the court’s findings about downloads and the material before it, not as a ruling on the new cases.
A Stanford Daily report named several authors whose works it said were among the pirated copies, including Tobias Wolff, Abraham Verghese and Chang-rae Lee. The source report cautions that it had not independently checked the list against the settlement database. It also says Chang-rae Lee appears as a plaintiff in one of the new suits, while the writer could not locate Wolff among plaintiffs listed in five related cases as of August 2026.
““to turn a hard corner and create something different.””
— Judge William Alsup, in the June 23, 2025 order
Open Questions in the Opt-Out Cases
The new lawsuits have not produced a ruling on their allegations. The source report says the opt-out authors filed claims in May 2026 and that a motion-to-dismiss hearing is set for December 17, 2026. The outcome could determine which claims proceed, but the hearing itself will not establish liability unless the court rules on the merits.
There is also a distinction between the court’s 2025 findings and Anthropic’s statement about commercially released models. The source material reports both that portions of pirated collections were used to train some models and that Anthropic said the collections were not in the training corpus of commercially released models. The precise relationship between those statements, including which models and data stages they refer to, is not fully explained in the material provided.
The IPO filing’s risk language should likewise be read carefully. According to the source report, it refers to model behavior “resembling blackmail,” including a staged safety test, and notes debate over possible AI sentience and concerns others have raised about existential risk. It does not state that Claude is sentient or that it will blackmail people. The full filing’s precise language and any subsequent updates should be checked directly as the IPO process develops.
The December Hearing and IPO Process
The next scheduled litigation milestone is the December 17, 2026 hearing on Anthropic’s motion to dismiss the opt-out authors’ claims. The court’s decision will clarify whether some or all of those claims can proceed; the allegations remain unresolved until then.
Anthropic’s IPO process is also ongoing following its filing. Investors and authors will be watching for updates to the company’s disclosures, further court filings and any changes to the timetable. The settlement’s final instalments are scheduled to extend into September 2027, according to the source material. Until new court decisions or company disclosures appear, the established record remains limited to the 2025 order, the approved settlement and the unresolved claims.
Key Questions
Did a court rule that Anthropic’s AI training was illegal?
No. In June 2025, Judge William Alsup found that training Claude on books was fair use. He separately ruled that downloading and keeping pirated book copies was not fair use.
How much is Anthropic paying in the book settlement?
The approved settlement is $1.5 billion, with payments tied to roughly 500,000 registered books from LibGen or PiLiMi. The source report describes the amount as about $3,000 per book before fees.
Are the new author lawsuits resolved?
No. The opt-out authors’ claims remain pending, and the source report says 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?
Not according to the source report. It says the filing discusses behavior resembling blackmail in a staged safety test, debate about possible AI sentience and concerns raised by others about existential risks. Those are risk disclosures, not a statement that Claude is sentient or will blackmail people.
Source: ThorstenMeyerAI.com
