🔍 Read the full analysis: How To Understand The Anthropic Book Piracy Allegations Ahead Of An IPO 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
Anthropic’s book-piracy dispute involves a 2025 court ruling, a $1.5 billion settlement approved in 2026, and new lawsuits that have not been decided. The judge found model training on books could be fair use but ruled that keeping millions of pirated copies was not; Anthropic disputes wrongdoing and says those collections were not used to train any commercially released model.
As Anthropic prepares for a possible public listing, its book-piracy dispute is again under scrutiny: a federal judge ruled in 2025 that training Claude on books could be fair use, but that downloading and keeping millions of pirated copies was not. A $1.5 billion settlement in a related case has been approved, while new claims from authors who opted out remain unresolved. The distinction matters because the court did not rule that all AI training on books was illegal.
The dispute began with Bartz v. Anthropic, a copyright case filed by three authors in 2024. In a June 23, 2025, summary-judgment order, U.S. District Judge William Alsup distinguished between how Anthropic obtained books and how it used them. He found that training Claude on books was fair use, describing the process as an effort “to turn a hard corner and create something different.” He also found that Anthropic’s purchase and scanning of print books was fair use.
The judge treated pirated copies differently. His order described more than seven million copies drawn from the Books3, Library Genesis and Pirate Library Mirror collections. It found that downloading and retaining pirated books was not fair use. The order accepted the authors’ account of the relevant facts for that pretrial decision; a jury was expected to consider remaining issues, including damages, but the case settled before trial.
The settlement, finally approved in July 2026, provides $1.5 billion for owners of about 500,000 registered books that Anthropic downloaded from Library Genesis or Pirate Library Mirror—about $3,000 per book before fees. It requires destruction of the downloaded collections and copies, subject to legal-hold obligations. The agreement covers specified past conduct through August 25, 2025, and does not release claims about Claude’s outputs. Anthropic denies wrongdoing and represented to the settlement court that neither collection, nor any portion of it, was in the training corpus of any commercially released model.
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 Legal Line Between Training and Theft
The distinction between training and acquisition may shape how courts handle other copyright cases involving generative AI. The ruling does not establish that every use of copyrighted books to train a model is lawful, nor does it mean that copying a work from an unauthorized source is excused if the material is later used for training. It addressed specific conduct and facts in one case.
The settlement also has financial and business significance. Its $1.5 billion payment resolves claims for covered books and authors, but it is not a licence for future training on copyrighted works and does not settle every claim related to Anthropic’s systems. For investors and readers, the key issue is the scope of the resolution and the risk left open by the cases still in court—not a claim that a judge found all Claude training unlawful.
Public descriptions of the case can conflate the court’s findings with the allegations in later lawsuits or with the settlement’s terms. Keeping those categories separate is particularly relevant as the company enters the public-markets process: its legal exposure depends on which claims remain, what evidence supports them and how courts rule.
Popular right now
LG C2 Series 65-Inch OLED evo 4K Smart TV (OLED65C2PUA, 2022)
Ninja CREAMi 7-in-1 Ice Cream Maker, NC301
Dyson Car+Boat Cordless Handheld Vacuum
As an affiliate, we earn on qualifying purchases.
In 2025, Judge Alsup’s order separated three activities: training on books, scanning copies Anthropic had bought, and downloading pirated collections for permanent storage. He found the first two fair use on the record before him and the third unlawful. His order said buying a copy later would not erase liability for the earlier theft. Because the matter settled, the jury did not reach a damages verdict on that conduct.
The settlement covers registered books in the specified collections, but some authors did not participate or opted out. They filed new lawsuits in 2026; those allegations have not been adjudicated. A Stanford Daily report named several Stanford-affiliated writers and said ten Tobias Wolff works were affected. That list should be treated as the paper’s reporting: the article’s headline and description, as recounted by ThorstenMeyerAI.com, risked conflating the June 2025 ruling with a separate May 2026 lawsuit. The source also said it could not verify Wolff’s name among plaintiffs in five related cases listed by August 2026, while Chang-rae Lee was named in one case.
Anthropic’s IPO filing is a separate document from the copyright rulings. Viral descriptions that it says Claude could “blackmail people” or is sentient are not faithful summaries of the filing, according to the source article. It says the filing discusses model behavior resembling blackmail in a staged safety test, possible regulation linked to debate about AI sentience, and concerns raised by others about existential risks. Those disclosures do not amount to a finding that Claude is sentient or a prediction that it will blackmail users.
“to turn a hard corner and create something different”
— Judge William Alsup, in the June 23, 2025, order
Claims Still Awaiting a Court Ruling
The new cases brought by authors who did not join the settlement remain allegations, not findings. The source says a hearing on Anthropic’s motion to dismiss is set for December 17, 2026; the outcome and any later trial schedule are not known. The precise set of plaintiffs and books may also change as cases proceed or filings are amended.
There is also a distinction between the 2025 order’s account of pirated copies being used to train some models and Anthropic’s settlement representation that the collections were not in the training corpus of any commercially released model. The source material does not provide enough detail to reconcile those statements across model versions, training stages or the meaning of “commercially released.” The settlement did not resolve every factual question about model training, nor did it decide claims concerning outputs.
The settlement payment schedule and the final amounts distributed to individual rights holders are also separate from the headline figure. The available source gives instalment dates and an approximate per-book amount before fees, but does not establish how much each claimant ultimately receives.
The December Hearing and IPO Filing
The next court milestone identified in the source is the December 17, 2026, hearing on Anthropic’s motion to dismiss the new authors’ claims. A decision could narrow or allow those cases to proceed, but no outcome should be assumed in advance. Later filings may clarify which books and plaintiffs are covered and what claims remain outside the settlement.
Anthropic’s IPO filing puts its disclosures under investor scrutiny, but the filing does not itself decide the copyright cases. Readers should distinguish the company’s stated position, the settlement terms, and the court’s 2025 findings. Any future court rulings or amendments to the IPO documents could add information; as of October 28, 2026, the new lawsuits have not produced a ruling on their merits.
Key Questions
Did a judge rule that Anthropic’s AI training was illegal?
No. In the June 2025 Bartz order, Judge Alsup found that training Claude on books was fair use on the facts before him. He separately ruled that downloading and keeping pirated copies was not fair use.
What does the $1.5 billion settlement cover?
It covers specified past claims involving about 500,000 registered books downloaded from Library Genesis or Pirate Library Mirror, with an approximate payment of $3,000 per book before fees. The agreement does not grant a licence for future training and does not release claims about Claude’s outputs.
Have the authors’ newer lawsuits been decided?
No. The new cases remain unresolved. The source reports a hearing on Anthropic’s motion to dismiss scheduled for December 17, 2026.
Does Anthropic’s IPO filing say Claude is sentient or will blackmail people?
Not as a confirmed fact. According to the source article, the filing discusses staged-test behavior resembling blackmail, debate over AI sentience and existential concerns raised by others. Those disclosures do not establish that Claude is sentient or will blackmail users.
Source: ThorstenMeyerAI.com
