The $1.5B AI Copyright Case: Anthropic Settlement and Author Payouts

The Anthropic copyright settlement is a $1.5 billion class-action deal — the largest publicly reported copyright recovery in United States history — resolving claims that Anthropic PBC trained its Claude chatbot on hundreds of thousands of books downloaded from pirate websites. Under the agreement, Anthropic will pay roughly $3,000 per eligible work and destroy the pirated files. A federal judge granted preliminary approval in September 2025, and final approval was still pending as of mid-2026.¹

What Anthropic Was Accused of Doing

Authors Andrea Bartz, Charles Graeber, and Kirk Wallace Johnson filed the class action on August 19, 2024, in the U.S. District Court for the Northern District of California.² The complaint alleged that Anthropic had downloaded more than seven million digitized books it knew were pirated and used them to train Claude.³

Three sources fed that library. In early 2021, Anthropic pulled roughly 196,640 books from a dataset called Books3. By mid-2021, it had downloaded at least five million copies from Library Genesis (LibGen). In July 2022, it grabbed at least two million more from Pirate Library Mirror (PiLiMi).⁴ Both LibGen and PiLiMi are shadow libraries that host copyrighted material without authorization.

The Ruling That Forced the Settlement

On June 23, 2025, Judge William Alsup split the case in two on summary judgment. Training an AI model on copyrighted books, he held, is fair use — he called the process “transformative — spectacularly so.” Anthropic’s models, in his view, “trained upon works not to race ahead and replicate or supplant them — but to turn a hard corner and create something different.”⁵

The pirated downloads were treated differently. Building a permanent library from LibGen, PiLiMi, and Books3 got no fair-use protection. “Such piracy of otherwise available copies is inherently, irredeemably infringing,” Alsup wrote, “even if the pirated copies are immediately used for the transformative use and immediately discarded.”⁵ With statutory damages under 17 U.S.C. § 504(c) reaching up to $150,000 per work for willful infringement, Anthropic’s exposure could have exceeded $15 billion at standard rates and reached $75 billion if a court found the conduct willful.⁶ That arithmetic pushed the parties to the table.

What the Settlement Pays and Requires

Anthropic agreed in late August 2025 to pay a minimum of $1.5 billion plus interest, and Judge Alsup granted preliminary approval on September 25, 2025.¹ The key terms:

Approximately $3,000 per eligible work, based on an anticipated class of about 500,000 titles. If more eligible works surface, Anthropic pays the same rate for each. The fund is non-reversionary, so nothing goes back to Anthropic.⁷
Anthropic must destroy all original files and derivative copies of works downloaded from LibGen and PiLiMi.
The release covers only past downloading and use of the pirated works. It is not a license for future training, and it preserves claims based on infringing AI outputs.¹
Rightsholders who decline to participate may pursue individual litigation.

The class covers legal and beneficial copyright owners of books downloaded by Anthropic from LibGen or PiLiMi, provided the works had an ISBN or ASIN and were registered with the U.S. Copyright Office within five years of publication.⁸ Books that were pirated but not timely registered fall outside the class. Publisher Macmillan offered to compensate its affected authors for the lost settlement award, and the Authors Guild urged other publishers to do the same.⁹

How the Money Reaches Authors

Several deductions come out of the fund before any check is cut. Class counsel sought 15% of the settlement, or $225 million, in attorneys’ fees. A separate group of law firms sought another $75 million — a request Judge Alsup sharply criticized in a December 2025 memorandum, questioning whether those “interloper” firms deserved any award at all.¹⁰ Administrative costs go to settlement administrator JND Legal Administration, and each of the three named plaintiffs is set to receive a $50,000 service award.¹¹

What is left gets divided equally among all eligible works for which a valid claim was filed. The estimated per-work payout is roughly $3,100.¹¹

How that per-work amount gets divided between author and publisher depends on the book. For trade and university press titles, the default is a 50/50 split between the author and publisher sides, though parties can agree to different terms based on their contracts. Self-published authors and those whose rights have reverted keep the full amount. Educational works like textbooks have no standard default, so claimants provide a good-faith estimate of the contractual split. If an author and publisher can’t agree, a court-appointed special master makes a binding decision.⁹

Claims, Opt-Outs, and Objections

The claims deadline was March 30, 2026, and submissions ran through anthropiccopyrightsettlement.com, administered by JND.¹² Only books on the court-approved Works List were eligible, and the court declined to reopen the list after it was finalized. Participation was high: class counsel reported filings for 440,490 of the 482,460 eligible works, a claim rate of 91.3%.¹³

After Judge Alsup transitioned to inactive status, the case was reassigned to Judge Araceli Martínez-Olguín, and the opt-out and objection deadline was extended to January 29, 2026.¹⁰ By the final approval hearing, 53 formal objections had been filed and 350 class members had opted out.¹⁴

The objections raised several themes. One argued that group copyright registrations covering dozens of novels were being counted as a single claimable work. Another challenged the exclusion of works published under pen names. A third contended that a one-time payment was inadequate given Anthropic’s ongoing profits from the training data. Attorney James H. Bartolomei III asked the court to reopen the opt-out window, arguing that key documents had been posted late to the settlement website.

The most notable opt-out was journalist John Carreyrou, author of Bad Blood. In December 2025, Carreyrou and five other authors filed a separate lawsuit against Anthropic, OpenAI, Google, Meta, xAI, and Perplexity AI, alleging downloads from shadow libraries including LibGen, Z-Library, and OceanofPDF.¹⁵ They argued that $3,000 per work amounts to “pennies on the dollar” compared to the $150,000 per work a jury could award for willful infringement.¹⁶ In June 2026, a federal judge ordered the claims split into separate lawsuits against each defendant.¹⁷

Where Things Stand Now

Judge Martínez-Olguín held the final approval hearing on June 3, 2026, and took the matter under submission. No ruling had been issued as of mid-2026. The settlement administrator has estimated that initial payments could begin as early as August 2026 if the court approves the deal and no appeals intervene.¹³

What the Settlement Means for the Rest of the AI Industry

Plaintiffs’ co-lead counsel Justin Nelson of Susman Godfrey called the $1.5 billion figure one that “far surpasses any other known copyright recovery.”¹ Even so, it is the first major payout tied specifically to AI use of creative works, and other cases are lining up behind it.

Plaintiffs’ lawyers in other cases have adopted what commentators call the “Bartz roadmap,” focusing not on whether AI training is fair use but on whether the training data was pirated. That sidesteps the still-unresolved fair-use debate. Music publishers have moved to amend their complaint in Concord Music Group v. Anthropic to add piracy claims modeled on Bartz.¹⁸

The fair-use half of Alsup’s ruling has not been universally followed. One day after Bartz, Judge Vince Chhabria in Kadrey v. Meta found that Meta’s use of copyrighted and pirated works to train its Llama model was fair use, rejecting Alsup’s approach of treating downloading and training as separate acts. Chhabria cautioned that his ruling was “solely based on the record before the court” and that the plaintiffs had failed to present evidence of market harm.¹⁹ Both decisions are district-court-level and not binding on other courts.²⁰

Because the Bartz settlement preserves claims based on infringing outputs, a separate question is being tested in Disney Enterprises, Inc. v. Midjourney Inc., filed in June 2025 in the Central District of California. Disney, NBCUniversal, and DreamWorks allege that Midjourney’s image generator reliably produces recognizable versions of copyrighted characters without authorization.²¹ The case was consolidated with a Warner Bros. lawsuit in November 2025 and is in discovery.²²

Analysts expect the $3,000-per-work figure — roughly four times the statutory damages floor — to serve as a baseline in future negotiations between AI companies and content owners, accelerating a shift toward direct licensing with authors, publishers, and collective rights organizations.²³²⁴ For smaller AI startups, commentators warn that the settlement raises the cost of entry, pushing new companies to secure content licenses upfront rather than risk the kind of bet-the-company litigation that forced Anthropic to settle.⁶

1
Susman Godfrey. Susman Godfrey Secures $1.5 Billion Settlement in Landmark AI Piracy Case
2
CourtListener. Bartz v. Anthropic PBC, No. 3:24-cv-05417
3
NPR. Anthropic Authors Settlement Pirated Chatbot Training Material
4
Publishers Weekly. Federal Judge Rules AI Training Is Fair Use in Anthropic Copyright Case
5
Goodwin Law. District Court Issues AI Fair Use Decision
6
IP Law Group. Anthropic’s $1.5B Settlement: A Landmark in the Evolving Copyright Terrain
7
ClassAction.org. Bartz et al. v. Anthropic PBC Notice
8
Authors Alliance. Bartz v. Anthropic: Judge Alsup Certifies Class for Rightsholders of 7 Million Books Used by Anthropic
9
Authors Guild. What Authors Need to Know About the Anthropic Settlement
10
Authors Alliance. Bartz v. Anthropic: Updated Opt-Out and Objection Dates and a New Judge
11
Wolters Kluwer Copyright Blog. The Bartz v. Anthropic Settlement: Understanding America’s Largest Copyright Settlement
12
AnthropicCopyrightSettlement.com. Anthropic Copyright Settlement
13
Society of Authors. Anthropic List of Stolen Works Published
14
Courthouse News Service. Authors, Publishers Near Final Approval of $1.5 Billion Anthropic Copyright Settlement
15
Publishers Weekly. Authors File New Lawsuit Against AI Companies Seeking More Money
16
Bloomberg Law. OpenAI, Anthropic, xAI Hit With Copyright Lawsuit From Writers
17
Law360. Anthropic, Other Tech Giants Get Authors’ Copyright Suit Split
18
Norton Rose Fulbright. AI in Litigation Series: An Update on AI Copyright Cases in 2026
19
Debevoise & Plimpton. Anthropic and Meta Decisions on Fair Use
20
Reed Smith. A New Look: Fair Use, Anthropic, Meta, Copyright, AI Training
21
Georgetown Law Tech Institute. Disney, NBC Universal, and DreamWorks File Major IP Lawsuit Against AI Image Generator Midjourney
22
ForensisGroup. Disney and Universal v. Midjourney: U.S. Generative AI Copyright Litigation Over Image Training and Outputs
23
Ropes & Gray. Anthropic’s Landmark Copyright Settlement: Implications for AI Developers and Enterprise Users
24
Buchanan Ingersoll & Rooney. Anthropic’s Copyright Settlement: Lessons for AI Developers and Deployers