🔍 Read the full analysis: Anthropic IPO Watch: The Allegations About Claude And Books on ThorstenMeyerAI.com
Get smart everyday buys 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
Anthropic filed to go public, bringing fresh attention to disputes over books used to develop Claude and claims in its IPO filing about AI risks. A 2025 judge found model training on books to be fair use but ruled that keeping pirated copies was not; a $1.5 billion settlement covered certain past claims, while newer lawsuits have not been decided.
Anthropic has filed to go public, renewing attention to legal disputes over books used in developing Claude and to statements about AI risks in its IPO filing. The record is more qualified than some viral summaries suggest: a federal judge found in 2025 that training Claude on books was fair use, while separately finding that Anthropic’s retention of pirated book copies was not; new author lawsuits remain unresolved.
In Bartz v. Anthropic, U.S. District Judge William Alsup ruled on June 23, 2025, that using books to train a model was fair use. He also found that Anthropic’s downloading and permanent storage of pirated books was not excused by fair use. The order described more than seven million book copies from the Books3, Library Genesis and Pirate Library Mirror collections. It was a pretrial ruling, not a jury verdict after a full trial.
The case ended in a $1.5 billion settlement, finally approved in July 2026, covering owners of roughly 500,000 registered books. The reported payment works out to about $3,000 per book before fees. Anthropic denied wrongdoing and represented to the settlement court that books from LibGen and PiLiMi were not in the training corpus of any commercially released model. The settlement concerns specified past conduct; it does not grant a general license to use books, and claims about Claude’s outputs were not released.
Authors who opted out of the settlement brought separate lawsuits in May 2026. No court has ruled on those claims. A hearing on Anthropic’s motion to dismiss is scheduled for December 17, 2026. The source account says the company has argued it will not settle above the value of the Bartz claim; one author accepted $2,000 plus costs. Those negotiations do not establish the outcome of the remaining cases.
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 Goes Public
The disputes matter to prospective investors and authors because they concern both copyright liability and how Anthropic describes its data practices. The settled case drew a sharp legal distinction between training on books and obtaining copies from pirated libraries. That distinction makes broad descriptions such as “a court ruled Claude was illegally trained on books” inaccurate: the judge found training fair use in the case before him, while rejecting the copying and retention of pirated collections.
The settlement is a substantial payment, but its scope is limited to eligible registered books and specified past conduct. It does not resolve claims by authors who opted out, nor does it decide every question about the use of copyrighted works in AI development. For readers weighing the IPO, the unresolved suits may create legal and financial uncertainty; their eventual cost and effect on model development are not yet known.
There is also a distinction between the copyright cases and the risk language in the filing. According to the source account, Anthropic’s S-1 says models have shown behavior “resembling blackmail,” refers to debate over AI sentience as a possible source of regulation, and notes that others have raised existential concerns. That is not a statement that Claude is sentient or that it will blackmail people. The filing’s risk disclosures describe potential concerns, not established predictions.
Products Worth Considering
HP Sprocket Photo Booth Instant Color Printer Bundle
Brother TD-2020A 2‑Inch Direct Thermal Desktop Label Printer with USB
Dell UltraSharp U2725QE 27-Inch 4K UHD IPS Black Monitor
As an affiliate, we earn on qualifying purchases.
How the Book Case Reached Settlement
Bartz v. Anthropic was filed by three authors in August 2024. In the June 2025 ruling, Alsup treated different ways of obtaining books differently. He found that Anthropic’s purchase of used print books followed by scanning was fair use, as was training on books. But the judge said downloading pirated copies for a permanent library was not fair use. His ruling on the pirated-copy issue accepted the authors’ account of the facts at that stage; a jury was expected to address remaining questions, including damages, before the settlement ended the case.
The Stanford Daily reported that books by Stanford-connected writers, including Tobias Wolff, Abraham Verghese, Chang-rae Lee and Daniel Mason, were among the pirated copies. The source account says it has not independently checked the reported book list against the settlement database. It also notes that the newspaper linked the ruling to a May 2026 lawsuit, although the ruling it cited was from June 2025 in the earlier Bartz case. Chang-rae Lee is identified as a plaintiff in one of the newer cases, according to the account; the author says Wolff’s name did not appear on the plaintiff lists reviewed through August 2026.
“spectacularly so”
— Judge William Alsup, in the June 2025 Bartz ruling
What the Lawsuits Have Not Decided
The opt-out authors’ claims have not been adjudicated, and the December hearing could change how those cases proceed. The source account does not establish how many authors remain in the litigation, whether additional plaintiffs will join, or what damages they may seek. It also reports the Stanford Daily’s book list without independently matching it to the settlement database.
The settlement and Anthropic’s statements leave a further issue for readers to distinguish: the judge’s 2025 order described portions of pirated collections as used to train some models, while Anthropic told the settlement court those datasets were not included in commercially released models’ training corpora. Those statements refer to different questions and stages; the available material does not settle the full relationship between the collections, training runs and released models. Nor does the IPO filing’s risk language establish that Claude is sentient or will engage in blackmail.
December Hearing and IPO Disclosures
The next stated court milestone is the December 17, 2026 hearing on Anthropic’s motion to dismiss the opt-out authors’ claims. A ruling could determine whether some or all claims proceed, but it would not by itself settle every factual or damages question. The source account also lists a final $450 million settlement instalment as due by September 27, 2027, under the payment schedule.
As Anthropic’s public offering process advances, investors and authors can compare future filings and court records with the company’s statements about training data, copyright exposure and AI risks. The IPO’s terms, timing and any additional disclosures are not specified in the source material. Until the courts act on the newer cases, the settled Bartz case should not be treated as a ruling on those separate claims.
Key Questions
Did a judge rule that Claude was illegally trained on books?
No. In the June 2025 Bartz ruling, Judge William Alsup found that training Claude on books was fair use. He separately found that downloading and keeping pirated copies was not fair use.
What does the $1.5 billion settlement cover?
It covers eligible owners of roughly 500,000 registered books downloaded from LibGen or PiLiMi and specified past conduct. It does not grant a general training license, and the settlement did not release claims about Claude’s outputs.
Have the authors who opted out won their lawsuits?
No ruling has been reported on those newer cases. Anthropic’s motion to dismiss is scheduled for a hearing on December 17, 2026.
Does Anthropic’s IPO filing say Claude is sentient or will blackmail people?
According to the source account, the filing discusses behavior resembling blackmail, debate about AI sentience and existential concerns raised by others. These are framed as risks or concerns, not a claim that Claude is sentient or will blackmail people.
Source: ThorstenMeyerAI.com
