Book authors settle copyright lawsuit with AI company Anthropic

Anthropic Reaches Historic Copyright Settlement with Authors in Landmark AI Case

Bottom Line Up Front

Three authors have settled what could have been the largest copyright class action ever certified—potentially representing up to 7 million authors whose books were allegedly pirated by Anthropic for AI training. The settlement avoids a December trial that could have resulted in hundreds of billions in damages and establishes the first industry precedent for compensating creators whose work was used without permission. While financial terms remain undisclosed, the agreement likely signals a shift toward licensing frameworks that could fundamentally reshape how AI companies acquire training data.

---

SAN FRANCISCO — In a development that could reshape the artificial intelligence industry's approach to copyrighted content, Anthropic has reached a settlement agreement with book authors in a class-action copyright infringement lawsuit that alleged the AI company engaged in "large-scale theft" of literary works to train its Claude chatbot.

The settlement, announced Tuesday, represents the first resolution in a series of high-stakes cases that will determine whether AI companies can continue training their models on copyrighted material without permission—or must begin paying licensing fees that could fundamentally alter the economics of the industry.

Settlement Brings Swift Resolution

Both sides have "negotiated a proposed class settlement," according to a federal appeals court filing, with terms expected to be finalized next week. The judge has given the parties a September 5 deadline to file requests for preliminary approval of the settlement.

A lawyer for the authors, Justin Nelson, called it a "historic settlement will benefit all class members," while Anthropic declined to comment on the deal. The financial terms of the settlement have not been disclosed, but legal experts suggest it could set a precedent for how authors are compensated in future licensing deals.

"The authors' side will be reasonably happy with this result, because they've essentially forced Anthropic into making a settlement," Luke McDonagh, an associate professor of law at LSE, told Fortune. "This might be the first domino to fall."

The Original Allegations

The 2024 class action lawsuit was brought by authors Andrea Bartz, Charles Graeber, and Kirk Wallace Johnson, who alleged Anthropic used the contents of millions of digitized copyrighted books to train the large language models behind Claude, including at least two works by each plaintiff.

The trio alleged that Anthropic's practices amounted to "large-scale theft," claiming the San Francisco-based company "seeks to profit from strip-mining the human expression and ingenuity behind each one of those works." They argued that rather than obtaining permission and paying fair compensation, Anthropic simply pirated their creative works.

The case was the first certified class action against an AI company over the use of copyrighted materials, and the quick settlement—coming just one month after the judge ruled the case could proceed to trial as a class action—represents a significant victory for content creators.

Class Action Authority: How Three Authors Can Speak for Millions

The settlement's authority to bind potentially millions of authors stems from the federal court's certification of the largest copyright class action ever approved. Under Federal Rule of Civil Procedure 23, three individual plaintiffs gained the legal right to represent a class that could include up to 7 million authors whose books were allegedly downloaded from pirate libraries Library Genesis (LibGen) and Pirate Library Mirror (PiLiMi).

To achieve class certification, the plaintiffs had to demonstrate four key requirements: numerosity (the class is too large for individual lawsuits to be practical), commonality (shared legal questions affect all class members), typicality (the named plaintiffs' claims are representative of the broader class), and adequacy (the representatives can fairly protect all class members' interests).

Judge William Alsup found these requirements met, noting that with potentially 7 million affected works, individual litigation would be "impracticable" and that common questions about Anthropic's systematic downloading from pirate sites affected all class members similarly. The court appointed experienced class counsel from Lieff Cabraser Heimann & Bernstein and Susman Godfrey to represent the entire class.

Crucially, class members who meet specific criteria—including legal ownership of copyrights, ISBN/ASIN numbers, and timely copyright registration—are automatically included in the settlement unless they actively opt out. As the Authors Guild explains, "You do not need to do anything to be a member of the class," though affected authors are encouraged to provide contact information to ensure they receive notices about the settlement.

The class certification was based on Anthropic's systematic downloading of approximately 5 million books from LibGen and 2 million from PiLiMi, with the court finding that the companies' detailed metadata catalogs—containing ISBN numbers, titles, authors, and hash values—provide sufficient information to identify affected works and their copyright holders. This represents a departure from traditional copyright cases involving individual works, instead addressing industrial-scale copying of entire digital libraries.

Court Creates Complex Legal Landscape

Before the settlement, the case had already produced a landmark legal ruling that will likely influence AI copyright litigation for years to come. In June, Senior U.S. District Judge William Alsup issued the first federal court order to meaningfully apply fair use doctrine to AI training on copyrighted books.

Judge Alsup ruled that Anthropic's use of legally acquired books to train their AI model was acceptable under fair use doctrine, calling the AI system's distilling from thousands of written works "quintessentially transformative." He wrote: "Like any reader aspiring to be a writer, Anthropic's [AI large language models] trained upon works not to race ahead and replicate or supplant them—but to turn a hard corner and create something different."

However, the judge made a crucial distinction that likely drove the settlement negotiations. While training on legally acquired books qualified as fair use, Alsup found that Anthropic had "downloaded for free millions of copyrighted books in digital form from pirate sites on the internet" as part of its effort "to amass a central library of 'all the books in the world.'"

This use of pirated content could not be defended as fair use, the judge ruled, allowing the authors' claims to proceed to trial. The ruling potentially exposed Anthropic to massive damages—with copyright law allowing for statutory damages of up to $150,000 per infringed work and the court finding that Anthropic may have illegally downloaded as many as 7 million books.

Understanding Fair Use in the AI Era

The Anthropic case sits at the center of a broader legal battle over how decades-old copyright doctrine applies to revolutionary new technology. Fair use allows limited use of copyrighted works without permission, balancing creators' rights against the public interest in innovation and free expression.

Courts analyze four statutory factors, but in practice often emphasize whether the use is "transformative"—adding new meaning or purpose rather than merely substituting for the original work—and whether it harms the market for the original.

Digital Precedents Favor Transformation

The current AI fair use debate builds on landmark cases from the early internet era. In Authors Guild v. HathiTrust (2014), the Second Circuit established that "the creation of a full‐text searchable database is a quintessentially transformative use," allowing university libraries to digitize millions of books for research and accessibility purposes.

Similarly, in Authors Guild v. Google Books (2015), courts ruled that Google's digitization of books for search functionality constituted fair use because it was "highly transformative," enabling entirely new capabilities rather than replacing the original works.

AI developers cite these precedents, arguing their use of copyrighted material is similarly transformative for training purposes. The Library Copyright Alliance supports this position, stating that "based on well-established precedent, the ingestion of copyrighted works to create large language models or other AI training databases generally is a fair use."

Competing Court Interpretations

However, recent rulings reveal significant judicial disagreement about how fair use applies to AI training. Just days before the Anthropic ruling, District Judge Vince Chhabria sided with Meta in a separate case brought by 13 authors including Sarah Silverman and Ta-Nehisi Coates, but used markedly different legal reasoning.

While Chhabria also found that generative AI was "necessarily a transformative use," he focused his analysis on whether Meta's training impacted the market for the original works rather than the transformative nature of the process itself. Unlike Judge Alsup, Chhabria found that using pirated content did not automatically disqualify fair use protection.

In contrast, the first major AI copyright ruling to reject fair use came in Thomson Reuters v. Ross Intelligence, where a federal court found that using copyrighted legal materials to train a competing AI research tool was "commercial and not transformative" because it served the same purpose as the original works.

Copyright Office Weighs In

The U.S. Copyright Office's May 2025 report on AI and copyright concluded that "there will not be a single answer regarding whether the unauthorized use of copyright materials to train AI models is fair use." The office identified a spectrum from likely fair use (noncommercial research that doesn't reproduce original works) to likely infringement (copying expressive works from pirated sources to generate competing content).

Significantly, the report rejected AI companies' analogy to human learning, noting that "this analogy rests on the faulty premise that fair use is a defense for all acts if those acts are performed for the purpose of learning."

Industry-Wide Legal Battles Intensify

The Anthropic settlement comes amid an unprecedented wave of copyright litigation targeting major AI companies, with cases filed against all the top players including Google, OpenAI, Microsoft, and Meta. Plaintiffs range from individual artists and authors to large companies like Getty Images and the New York Times.

Music Industry Escalates Pressure

The music industry has emerged as particularly aggressive in challenging AI companies. Universal Music Group, Sony Music, and Warner Music Group, represented by the Recording Industry Association of America, have sued AI music generators Suno AI and Udio AI for alleged copyright infringement—the first lawsuit targeting audio-based generative AI.

Universal Music Group has also filed a separate lawsuit against Anthropic, alleging the company illegally trained its AI programs on copyrighted lyrics. Meanwhile, Meta and Universal have expanded their partnership to address unauthorized AI-generated music content, reflecting growing industry efforts to balance innovation with creator rights.

Global Publishers Fight Back

Japanese newspapers have emerged as major challengers to AI companies' data practices. Japan's largest newspaper, Yomiuri Shimbun, sued Perplexity AI in August, followed by Nikkei and Asahi filing a joint lawsuit seeking ¥2.2 billion ($15 million) in damages each.

These publishers allege that Perplexity ignored their robots.txt files designed to prevent automated content scraping and "repeatedly served up the content in response to user queries" without authorization. The cases represent the first major copyright challenges by Japanese media companies against AI firms.

Search Engine Copyright Battles

Perplexity AI faces mounting pressure from publishers worldwide. The company has received cease-and-desist letters from The New York Times, the BBC, and Condé Nast, while News Corp successfully fought off Perplexity's attempt to dismiss their copyright lawsuit on jurisdictional grounds.

The Perplexity cases focus on "retrieval-augmented generation" (RAG) technology, which differs from traditional AI training by accessing and copying content in real-time rather than during a one-time training process. Publishers allege this approach "repackages original copyrighted works into verbatim or near verbatim summaries."

Implications for AI Industry Future

The Anthropic settlement may signal a shift toward negotiated resolutions rather than prolonged litigation, potentially creating a framework for how AI companies compensate creators for training data.

"Some companies are already engaging in licensing content from publishers. So this may push us forward towards a situation in which the AI companies reach settlements that effectively retrospectively pay license fees for what they've already used," McDonagh said. "Perhaps [it's] the beginning of a licensing future for the AI industry, where authors can be paid for the use of their copyrighted works in the training of AI."

However, the path forward remains uncertain. AI developers have argued that mandatory licensing schemes would be "too expensive and cumbersome," potentially stifling innovation. Rights holders counter that licensing costs should be seen as "the price of doing business" and that the burden of obtaining permission should not excuse unauthorized use.

The stakes are enormous. As one legal observer noted, "The outcomes of these cases are set to have an enormous impact on the future of AI. In effect, they will decide whether or not model makers can continue ordering up a free lunch."

Legal scholars anticipate that "over the next few months and years, lawsuits will force courts to set new precedent in this area and draw the contours of copyright law as applied to generative AI." With dozens of similar cases pending and the Supreme Court potentially called upon to provide definitive guidance, the legal landscape for AI training on copyrighted content remains in flux.

The Anthropic settlement, while historic, represents just the beginning of a broader reckoning between the AI industry and content creators over the fundamental question of who owns and controls the information that powers artificial intelligence.

---

Sources

Primary Court Documents and Legal Sources:

1. Bartz v. Anthropic PBC, No. 3:23-cv-07024-WHO (N.D. Cal. 2025) - The landmark federal case establishing fair use precedent for AI training
2. Kadrey v. Meta Platforms, No. 3:23-cv-03417-VC (N.D. Cal. June 25, 2025) - Companion case with different fair use reasoning
3. Thomson Reuters Enterprise Centre GMBH v. Ross Intelligence Inc., No. 1:20-cv-00613-UNA (D. Del. Feb. 2025) - First major AI copyright ruling rejecting fair use
4. Authors Guild, Inc. v. HathiTrust, 755 F.3d 87 (2d Cir. 2014) - Foundational digital library fair use precedent
5. Authors Guild, Inc. v. Google, Inc., 804 F.3d 202 (2d Cir. 2015) - Google Books fair use victory
6. U.S. Copyright Office Report: "Copyright and Artificial Intelligence Report" (May 2025) - Official government position on AI training fair use

News Reports and Analysis:

7. NPR. "Federal judge rules in AI company Anthropic's favor in landmark copyright infringement lawsuit brought by authors." June 26, 2025. https://www.npr.org/2025/06/25/nx-s1-5445242/federal-rules-in-ai-companys-favor-in-landmark-copyright-infringement-lawsuit-authors-bartz-graeber-wallace-johnson-anthropic
8. North Carolina Lawyers Weekly. "Anthropic settles copyright lawsuit with book authors." August 27, 2025. https://nclawyersweekly.com/2025/08/27/anthropic-settles-copyright-lawsuit-authors/
9. Fortune. "Anthropic landmark copyright settlement with authors may set a precedent for the whole industry." August 27, 2025. https://fortune.com/2025/08/27/anthropic-settlement-with-authors-may-be-the-first-domino-to-fall-in-ai-copyright-battles/
10. Euronews. "Anthropic settles AI copyright lawsuit with book authors." August 27, 2025. https://www.euronews.com/next/2025/08/27/anthropic-settles-ai-copyright-lawsuit-with-book-authors
11. PYMNTS.com. "Anthropic and Authors Settle Copyright Infringement Lawsuit Targeting AI Training." August 26, 2025. https://www.pymnts.com/legal/2025/anthropic-and-authors-settle-copyright-infringement-lawsuit-targeting-ai-training/
12. ABC News. "Book authors settle copyright lawsuit with AI company Anthropic." August 26, 2025. https://abcnews.go.com/US/wireStory/book-authors-settle-copyright-lawsuit-ai-company-anthropic-125003388
13. U.S. News & World Report. "Book Authors Settle Copyright Lawsuit With AI Company Anthropic." August 26, 2025. https://www.usnews.com/news/business/articles/2025-08-26/book-authors-settle-copyright-lawsuit-with-ai-company-anthropic
14. Hartford Courant. "Book authors settle copyright lawsuit with AI company Anthropic." August 26, 2025. https://www.courant.com/2025/08/26/authors-anthropic-settlement/
15. Startup News. "Book authors settle copyright lawsuit with AI company Anthropic." August 27, 2025. https://startupnews.fyi/2025/08/27/book-authors-settle-copyright-lawsuit-with-ai-company-anthropic/
16. Insurance Journal. "Anthropic Settles Class Action From Authors Alleging Copyright Infringement." August 27, 2025. https://www.insurancejournal.com/news/national/2025/08/27/837098.htm

Legal Analysis and Commentary:

17. Electronic Frontier Foundation. "Two Courts Rule On Generative AI and Fair Use — One Gets It Right." July 11, 2025. https://www.eff.org/deeplinks/2025/06/two-courts-rule-generative-ai-and-fair-use-one-gets-it-right
18. Ballard Spahr. "Novel Ruling Offers Framework for 'Fair Use' of Copyrighted Material for Training AI Systems." July 2025. https://www.ballardspahr.com/insights/alerts-and-articles/2025/07/novel-ruling-offers-framework-for-fair-use-of-copyrighted-material-for-training-ai-systems
19. Davis+Gilbert LLP. "Court Rules AI Training on Copyrighted Works Is Not Fair Use — What It Means for Generative AI." February 27, 2025. https://www.dglaw.com/court-rules-ai-training-on-copyrighted-works-is-not-fair-use-what-it-means-for-generative-ai/
20. Skadden, Arps, Slate, Meagher & Flom LLP. "Copyright Office Weighs In on AI Training and Fair Use." May 2025. https://www.skadden.com/insights/publications/2025/05/copyright-office-report
21. Crowell & Moring LLP. "AI Companies Prevail in Path-Breaking Decisions on Fair Use." June 2025. https://www.crowell.com/en/insights/client-alerts/ai-companies-prevail-in-path-breaking-decisions-on-fair-use
22. Kluwer Copyright Blog. "Is Generative AI Fair Use of Copyright Works? NYT v. OpenAI." February 29, 2024. https://copyrightblog.kluweriplaw.com/2024/02/29/is-generative-ai-fair-use-of-copyright-works-nyt-v-openai/
23. Quinn Emanuel. "Lead Article: The Fair Use Frontier: Copyright Law in the Age of AI and Machine Learning." February 2, 2024. https://www.quinnemanuel.com/the-firm/publications/lead-article-the-fair-use-frontier-copyright-law-in-the-age-of-ai-and-machine-learning/
24. BitLaw. "Fair Use and the Training of AI Models on Copyrighted Works." 2025. https://www.bitlaw.com/ai/AI-training-fair-use.html
25. Stanford HAI. "Reexamining 'Fair Use' in the Age of AI." 2025. https://hai.stanford.edu/news/reexamining-fair-use-age-ai

Industry and Academic Sources:

26. Association of Research Libraries. "Training Generative AI Models on Copyrighted Works Is Fair Use." January 23, 2024. https://www.arl.org/blog/training-generative-ai-models-on-copyrighted-works-is-fair-use/
27. Association of Research Libraries. "Authors Guild v. HathiTrust Litigation Ends in Victory for Fair Use." April 16, 2019. https://www.arl.org/news/authors-guild-v-hathitrust-litigation-ends-in-victory-for-fair-use/
28. Electronic Frontier Foundation. "Authors Guild v. HathiTrust." March 9, 2018. https://www.eff.org/cases/authors-guild-v-hathitrust
29. Electronic Frontier Foundation. "Digitizing Books Is Fair Use: Author's Guild v. HathiTrust." October 11, 2012. https://www.eff.org/deeplinks/2012/10/authors-guild-vhathitrustdecision
30. The Media Institute. "Authors Guild v. HathiTrust: Fair Use Comes to Digital Book Repositories." August 6, 2014. https://www.mediainstitute.org/2014/08/06/authors-guild-v-hathitrust-fair-use-comes-to-digital-book-repositories/

Music Industry and Related Lawsuits:

31. Crowell & Moring LLP. "Major American Music Labels Sue Generative AI Music Platforms in First Case of Its Kind Over AI Audio." 2025. https://www.crowell.com/en/insights/client-alerts/major-american-music-labels-sue-generative-ai-music-platforms-in-first-case-of-its-kind-over-ai-audio
32. CoinGeek. "Universal Music Group files copyright infringement lawsuit vs Anthropic AI." June 5, 2025. https://coingeek.com/universal-music-group-files-copyright-infringement-lawsuit-vs-anthropic-ai/
33. Perplexity AI. "Meta and Universal Expand Deal to Tackle AI Music." 2025. https://www.perplexity.ai/page/meta-and-universal-expand-deal-UVp_uTtVT5aTN5vJEM21WQ

International and Perplexity Cases:

34. Nieman Journalism Lab. "Japan's largest newspaper, Yomiuri Shimbun, sues AI startup Perplexity for copyright violations." August 11, 2025. https://www.niemanlab.org/2025/08/japans-largest-newspaper-yomiuri-shimbun-sues-perplexity-for-copyright-violations/
35. The Register. "Asahi, Nikkei sue Perplexity AI for copyright infringement." August 26, 2025. https://www.theregister.com/2025/08/26/perplexity_asahi_nikkei_lawsuits

Comprehensive Legal Timelines:

36. Sustainable Tech Partner. "Generative AI Lawsuits Timeline: Legal Cases vs. OpenAI, Microsoft, Anthropic, Nvidia, Perplexity, Intel and More." August 25, 2025. https://sustainabletechpartner.com/topics/ai/generative-ai-lawsuit-timeline/
37. Copyright Alliance. "AI Lawsuit Developments in 2024: A Year in Review." May 14, 2025. https://copyrightalliance.org/ai-lawsuit-developments-2024-review/
38. MIT Technology Review. "What comes next for AI copyright lawsuits?" July 1, 2025. https://www.technologyreview.com/2025/07/01/1119486/ai-copyright-meta-anthropic/
39. Slate. "Why Sarah Silverman and Ta-Nehisi Coates sued Meta over copyright." June 30, 2025. https://slate.com/technology/2025/06/ai-copyright-lawsuits-anthropic-meta-openai-google.html

Legal Databases and Court Records:

40. Justia. "Authors Guild, Inc. v. HathiTrust, No. 12-4547 (2d Cir. 2014)." https://law.justia.com/cases/federal/appellate-courts/ca2/12-4547/12-4547-2014-06-10.html
41. Stanford Copyright and Fair Use Center. "Authors Guild, Inc. v. HathiTrust." January 31, 2023. https://fairuse.stanford.edu/case/authors-guild-inc-v-hathitrust/
42. U.S. Copyright Office. "Fair Use Index - Authors Guild, Inc. v. HathiTrust, 755 F.3d 87 (2d Cir. 2014)." https://www.copyright.gov/fair-use/summaries/authorsguild-hathitrust-2dcir2014.pdf
43. Davis Wright Tremaine. "2nd Circuit Issues Its Fair Use Ruling in Authors Guild v. HathiTrust." June 2014. https://www.dwt.com/insights/2014/06/2nd-circuit-issues-its-fair-use-ruling-in-authors
44. Wikipedia. "Authors Guild, Inc. v. Google, Inc." March 13, 2025. https://en.wikipedia.org/wiki/Authors_Guild,_Inc._v._Google,_Inc.
45. Wikipedia. "Authors Guild, Inc. v. HathiTrust." May 7, 2025. https://en.wikipedia.org/wiki/Authors_Guild,_Inc._v._HathiTrust
46. Book authors settle copyright lawsuit with AI company Anthropic