🔍 Read the full analysis: Claude’s Training Data Controversy: Claims, Evidence, And IPO Implications on ThorstenMeyerAI.com
Get hardware and tech essentials 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 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. Anthropic later agreed to a $1.5 billion settlement over claims tied to roughly 500,000 books; new lawsuits filed by authors in 2026 have not been decided. Its IPO filing discusses AI-related legal and safety risks, but the supplied material does not establish that the filing says Claude is sentient or will blackmail people.
Anthropic’s book-training dispute is back in focus as the company pursues an initial public offering and authors bring new lawsuits over copyrighted works. The record distinguishes a June 2025 ruling that found training Claude on books to be fair use from the court’s finding that downloading and retaining pirated copies was not; the newer claims have not yet been decided.
In Bartz v. Anthropic, U.S. District Judge William Alsup ruled on June 23, 2025 that training an AI model on books was fair use under the facts before him. He also found that Anthropic’s acquisition and scanning of books it had bought was fair use. Separately, he ruled against the company’s downloading and permanent storage of pirated books. The decision was a pretrial summary-judgment order, not a jury verdict on damages.
The court record described more than seven million book copies from collections including Books3, Library Genesis and the Pirate Library Mirror. Anthropic agreed to a $1.5 billion settlement covering claims over roughly 500,000 registered books. The settlement was finally approved in July 2026, according to the source material, and requires destruction of the downloaded collections and copies, subject to legal-hold obligations. It covers specified past conduct, not claims about Claude’s outputs.
Anthropic disputes wrongdoing and told the settlement court that the LibGen and PiLiMi datasets, or portions of them, were not in the training corpus of any commercially released model. The supplied account also says the court record showed portions of pirated collections were used to train some Anthropic models. Those statements concern different questions—what material was copied and what was used in commercially released models—and should not be collapsed into a claim that a court found every pirated book was used to train Claude.
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
Copyright Exposure as Anthropic Seeks an IPO
The distinction between training practices and data acquisition matters to authors, investors and other AI companies. The 2025 ruling did not declare all AI training lawful, nor did it excuse obtaining books through piracy. Instead, it separated the use of books for training from the way Anthropic obtained and stored some copies, leaving companies exposed to legal claims over how training material is sourced.
The settlement’s scale—$1.5 billion for about 500,000 books—shows the potential financial consequences of copyright disputes, though it does not settle the new cases or establish liability for every dataset. For prospective investors, unresolved litigation and the company’s disclosures about legal and regulatory risks are relevant to evaluating uncertainty. The source material does not establish that the dispute will prevent an IPO or determine its financial outcome.
Products Worth Considering
LG UltraGear 45GX900A-B 45-Inch Curved OLED Gaming Monitor
Logitech MX Vertical Wireless Ergonomic Mouse – Graphite
Logitech MX Mechanical Full-Size Wireless Mechanical Keyboard, Tactile Quiet Switches, Backlit Keys, Bluetooth, USB-C, Multi-Device, Graphit
As an affiliate, we earn on qualifying purchases.
What the 2025 Ruling Actually Decided
The case behind the 2025 order was brought by three authors in August 2024. Judge Alsup treated the purchased print books, which Anthropic cut apart and scanned, differently from copies downloaded from pirated collections and kept in a permanent library. On training, he described Claude’s process as creating something different from its source material and found fair use on the record before him.
The finding on piracy was also made at the pretrial stage: the judge accepted the authors’ version of the facts for that part of the motion. A jury was expected to address remaining issues, including damages, but the parties settled before trial. The settlement was finally approved in July 2026, according to the source material. Separately, authors filed new suits in May 2026; those allegations are not findings from the Bartz ruling.
Coverage by The Stanford Daily named Stanford-affiliated authors it said had works among the pirated copies. The source article cautions that it did not independently verify the reported book list against the settlement database. It also says the student paper’s description of a ruling in a May 2026 lawsuit is inaccurate: the linked ruling was from June 2025 and arose in a different case.
“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 its settlement representation
New Claims and Data Questions Remain
The 2026 lawsuits remain unresolved. No ruling was issued in May 2026, and a hearing on Anthropic’s motion to dismiss is set for December 17, 2026. The claims, the defendants’ responses and the court’s eventual assessment should not be confused with the findings in the 2025 case.
It is also unclear from the supplied material which individual books in the Stanford Daily’s account are registered for settlement payment. The article reports that Chang-rae Lee is a named plaintiff in one newer suit, but says it could not find Tobias Wolff among plaintiffs listed in five related cases as of August 2026; Wolff could be part of another action or a later amendment. The complete status of each author’s claim is not established here.
Viral descriptions of Anthropic’s IPO filing also require care. The source says the S-1 discusses models displaying behavior “resembling blackmail” in a safety-test setting, potential regulation related to debate over AI sentience, and existential concerns raised by others. That does not establish that the filing says Claude is sentient or will blackmail people. The supplied material does not provide enough detail to assess the full wording or context of every S-1 disclosure.
December Hearing and IPO Disclosures
The next identified court milestone is the December 17, 2026 hearing on Anthropic’s motion to dismiss the new author claims. The court could then allow some claims to proceed, dismiss them or take other steps; the outcome is not known. The cases may also clarify how the plaintiffs connect their specific works and alleged uses to the claims.
Anthropic is also pursuing an IPO, according to the source material. Investors and authors will be watching for the company’s disclosures about litigation, copyright practices and AI safety risks, as well as any developments in the pending cases. No IPO outcome or resolution of the new lawsuits is confirmed in the information provided.
Key Questions
Did a court rule that Anthropic illegally trained Claude on books?
Not in those terms. In June 2025, Judge William Alsup found that training on books was fair use on the record before him, but ruled that downloading and retaining pirated copies was not. The decision was not a final jury verdict on damages.
What does the $1.5 billion settlement cover?
According to the supplied material, it covers claims related to roughly 500,000 registered books downloaded from LibGen or PiLiMi and was finally approved in July 2026. It addresses specified past conduct and does not release claims about Claude’s outputs.
Have the authors’ newer lawsuits been decided?
No ruling on the new claims is reported. The source says authors filed lawsuits in May 2026 and 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?
The supplied account says no. It describes S-1 language about behavior “resembling blackmail” in a staged safety test, debate over AI sentience and existential concerns raised by others. Those points do not establish that Claude is sentient or that it will blackmail people.
Source: ThorstenMeyerAI.com
