- U.S. District Judge Araceli Martínez-Olguín approved the $1.5 billion settlement on July 21, calling it 'meaningful relief' for affected authors [1] - The payout covers roughly 500,000 books at approximately $3,000 per work, with 91% of the more than 482,000 eligible books already claimed [2] - Anthropic must destroy pirated files from Library Genesis and Pirate Library Mirror datasets within 30 days of final judgment [3] - Attorney fees were set at $101.5 million — roughly 7% of the fund — far less than the $187.5 million originally requested [4] - Because Anthropic settled rather than appealing, Judge William Alsup's earlier fair use ruling carries no binding precedent for the broader AI industry
[5] U.S. District Judge Araceli Martínez-Olguín on Monday granted final approval to a $1.5 billion class action settlement between Anthropic and thousands of authors who accused the AI company of training its Claude chatbot on pirated copies of their books. The deal, filed in the Northern District of California in San Francisco, is believed to be the largest copyright settlement in U.S. history [1].
The settlement covers more than 482,000 books at approximately $3,000 per work, with 91% already claimed by authors and publishers who are now due payment [2]. Anthropic will also be required to destroy the original files it downloaded from pirate websites — including Library Genesis and Pirate Library Mirror — within 30 days of the final judgment
.
[3]The case was first brought in 2024 by bestselling thriller novelist Andrea Bartz and two other authors. It is the first major U.S. copyright lawsuit against an AI developer to end in a payout, and arrives as Google, Meta, OpenAI, and Midjourney face similar litigation [5].
The Settlement #
Judge Martínez-Olguín rejected objections from authors who argued the $3,000 per-book amount was too low, writing that the complaints were 'not grounded in a realistic assessment of the overall risks and rewards' of continued litigation [4]. Lead plaintiff attorney Justin Nelson called it 'the largest known copyright recovery in history'
.
[4]Attorney fees were awarded at $101.5 million, or roughly 7% of the total settlement fund, along with approximately $2.6 million in expenses and an anticipated $18.2 million in additional costs. The fee was substantially less than the $187.5 million plaintiffs' counsel had initially sought [4].
Some authors and publishers opted out of the class and are continuing separate cases against Anthropic [5].
How the Case Unfolded #
The lawsuit originated in 2024 when Bartz and two co-plaintiffs alleged Anthropic had used their copyrighted works without permission or compensation to train its large language models. The case landed before U.S. District Judge William Alsup, who issued a pivotal ruling in June 2025 finding that training AI models on copyrighted text constitutes fair use under copyright law [5].
However, Alsup also found that Anthropic had illegally obtained a portion of its training library by down more than 7 million books from pirate websites, including Library Genesis and Pirate Library Mirror, rather than purchasing them [2]. That distinction — between lawful fair-use training and unlawful acquisition of copyrighted material — became the legal basis for the settlement
.
[5]Judge Alsup granted preliminary approval to the $1.5 billion deal in September 2025 before the case was reassigned to Judge Martínez-Olguín for final approval [1].
Why It Matters for the AI Industry #
The settlement closes the first major copyright case against an AI company but leaves critical legal questions unanswered. Because Anthropic chose to settle rather than appeal Alsup's fair use ruling, that determination will never be reviewed by a higher court and carries no binding precedent across the industry [5].
That creates an unusual legal landscape: AI companies can point to Alsup's ruling as supportive of fair use in training, while authors and publishers can argue it was never tested on appeal. The question of whether AI training on copyrighted material is lawful remains open for future cases [5].
The broader litigation pipeline is growing. Publishers Hachette, Cengage, and Elsevier recently joined author Scott Turow and advocacy group S.C.R.I.B.E. in bringing a fresh class action against Google over its Gemini AI training practices. OpenAI, Meta, and Midjourney all face their own pending copyright suits [5].
What's Next #
Anthropic has 30 days from the final judgment to destroy the pirated book files and certify their deletion [3]. Payments to authors and publishers who have filed claims will follow, with the timeline for disbursement still to be determined.
The outcome is likely to influence settlement negotiations in the parallel cases against other AI companies. The $3,000-per-book benchmark, while criticized by some authors as inadequate, now serves as a concrete data point for how courts and parties value copyrighted works used in AI training [4].
Companies mentioned #
Further sources #
[[1] ABC News / AP, 'Judge approves a $1.5B Anthropic settlement over pirated books … ↗](https://abcnews.com/Technology/wireStory/judge-approves-15b-anthropic-settlement-pirated-books-train-134949964)
[[2] ABC News / AP, 'Judge approves a $1.5B Anthropic settlement over pirated books … ↗](https://abcnews.com/Technology/wireStory/judge-approves-15b-anthropic-settlement-pirated-books-train-134949964)
[[3] The Verge, 'Anthropic's $1.5 billion book piracy settlement approved by judge,'… ↗](https://www.theverge.com/ai-artificial-intelligence/968724/anthropic-authors-settlement-ai-copyright-approved)
[4] Claims Journal, 'Judge Approves Anthropic's $1.5B Settlement of Copyright Lawsu… ↗
[5] Yahoo Finance / Reuters, 'Anthropic $1.5 billion copyright settlement approved … ↗ The stories that matter, in one email. Free — unsubscribe anytime.