The Bartz v. Anthropic settlement is a $1.5 billion class action resolution in which AI company Anthropic agreed to pay copyright holders whose books it downloaded from pirate sites to build its training library. Filed in August 2024 in the U.S. District Court for the Northern District of California, the case covers roughly 482,460 eligible works and is the largest copyright class action settlement in American history. It cleared its final fairness hearing in May 2026 and is awaiting a ruling on final approval, with payments expected to begin in late fall 2026 at the earliest.
What Anthropic Did
Between early 2021 and mid-2022, Anthropic assembled a permanent research library out of pirated ebooks. Cofounder Ben Mann pulled about 196,640 books from Books3, a repository he knew contained pirated copies. In June 2021, he used BitTorrent to mass-download at least five million ebook files from Library Genesis (LibGen). In July 2022, Anthropic downloaded at least two million more from the Pirate Library Mirror (PiLiMi), filling gaps in what it already had.
Anthropic kept these copies in a central library even when specific titles weren’t used to train Claude, its large language model. Judge William Alsup, who presided over the case before his retirement, noted that CEO Dario Amodei had acknowledged the company “preferred to steal” the books rather than deal with the “legal/practice/business slog” of buying them legitimately.
The Ruling That Forced the Settlement
On June 23, 2025, Judge Alsup split the case in two. He ruled that using copyrighted books to train AI models is “exceedingly transformative” and qualifies as fair use under Section 107 of the Copyright Act, and that digitizing books Anthropic had lawfully bought was also fair use. But he ruled that Anthropic’s creation of a “permanent, general-purpose library” from millions of stolen books was not fair use and was “inherently, irredeemably infringing.” The court rejected the argument that a research purpose justified the downloads.
That split is what drove the settlement. Statutory damages for willful infringement reach up to $150,000 per work, and with roughly 482,460 eligible works, Anthropic faced potential liability as high as $72 billion. The company chose to settle.
What the Settlement Pays and Requires
Anthropic agreed to pay a minimum of $1.5 billion plus statutory interest in four installments. The first $300 million was due within a week of preliminary approval, which Judge Alsup granted on September 25, 2025. A second $300 million payment is due within a week of final approval. Two more payments of $450 million each are scheduled for September 2026 and September 2027.
The deal also requires Anthropic to destroy the original files it downloaded from LibGen and PiLiMi, along with any copies derived from them, and to certify the destruction in writing. It releases Anthropic only for conduct occurring before August 25, 2025. It does not license Anthropic to train on these works going forward, and it explicitly preserves claims based on infringing outputs generated by Claude.
Whose Books Qualify
The class, certified in August 2025, covers copyright owners of books in the LibGen or PiLiMi datasets that Anthropic downloaded, subject to filters. Each eligible book must have an ISBN or ASIN. Its copyright must have been registered with the U.S. Copyright Office within five years of publication and either before Anthropic’s download date or within three months of publication.
About 482,460 works passed those filters. Eligible claimants include authors, publishers, estates, and any entity holding reproduction rights. Foreign rightsholders qualify if their works carry U.S. copyright registrations. Books from the Books3 dataset were excluded because that collection lacked the metadata needed to identify individual titles and authors. Books Anthropic bought and scanned lawfully were also excluded.
Three authors served as class representatives: Andrea Bartz, Charles Graeber, and Kirk Wallace Johnson. Each received a $50,000 service award.
How Payouts Work
The fund is distributed equally among valid claims on a per-work basis. The initial estimate was roughly $3,000 per title, though the precise figure depends on total claims, legal fees, and accrued interest. As of April 2026, with 440,490 works claimed out of 482,460 eligible titles, the estimated base payout was approximately $2,932 per work before interest.
When both an author and publisher hold rights to the same book, the default allocation is a 50/50 split. Claimants could request a different split by submitting their publishing contract. Self-published authors and those whose rights had reverted receive 100% of the award for their titles. Educational works have no default split, and claimants had to give a good-faith estimate based on their contracts.
Key Dates
The opt-out deadline was January 29, 2026. The claims deadline was March 30, 2026. By the final fairness hearing on May 14, 2026, the claim rate had reached 92.77%, with 447,576 works claimed. Only 350 class members opted out, and 53 filed objections.
Judge Alsup moved to inactive status in late 2025, and the case was reassigned to Judge Araceli Martínez-Olguín, who presided over the fairness hearing. Seven people presented objections in person. About half of the 53 written objections came from people asking to have their works added to the settlement rather than challenging its terms. Other objections concerned notice sufficiency, pseudonymous publications, the adequacy of the payout, and the exclusion of foreign works without U.S. registrations.
Judge Martínez-Olguín did not approve the settlement from the bench. She ordered Anthropic to file a brief on why five untimely opt-out requests should not be honored. As of mid-June 2026, final approval remains pending, though observers at the hearing reported that nothing appeared likely to derail it. Payments are expected to begin in late fall 2026 at the earliest, arriving in installments as Anthropic pays into the fund.
Opting Out and Its Risks
Author John Carreyrou and five other writers opted out and filed individual copyright infringement suits in December 2025. Their case, Carreyrou et al. v. Anthropic PBC et al., names six AI companies as defendants: Anthropic, OpenAI, Google, Meta, xAI, and Perplexity AI. They argued that $3,000 per work amounted to “pennies on the dollar” and are seeking up to $150,000 per work per defendant, as much as $900,000 per title.
The Authors Guild cautioned members thinking about similar moves. While acknowledging the $3,000 figure feels “paltry,” the Guild noted that courts rarely award maximum statutory damages, and that the median award in cases involving more than twelve works is itself around $3,000 per work. The Guild also warned that individual litigation is expensive, slow, and risky, with costs potentially exceeding $1 million before any contingency fee.
When Publishers Failed to Register
Some authors discovered their books were ineligible because a publisher had contractually agreed to register the copyright and never did. Because the settlement required timely U.S. Copyright Office registration, these authors lost out through no fault of their own. The Authors Guild praised Macmillan for offering to compensate affected authors for their lost settlement awards and publicly urged other publishers to do the same.
What This Settlement Signals
The case drew a distinction now shaping AI copyright litigation: what AI companies do with copyrighted material during training may qualify as fair use, but how they acquire it matters enormously. Judge Alsup’s ruling that training is transformative gave AI developers a legal foothold. His finding that pirating the source material is infringing gave copyright holders a clear path to recovery when companies cut corners on acquisition.
The $3,000-per-work benchmark, roughly four times the $750 statutory minimum, has become a reference point in ongoing negotiations. Analysts expect future plaintiffs to treat it as a floor. The requirement that Anthropic identify, segregate, and destroy pirated datasets is likely to become a standard demand in similar cases. More than 75 AI copyright lawsuits had been filed by early 2026, with major consolidated actions pending against OpenAI, Google, and Meta.