Authors have until March 30 to claim their share of the largest copyright settlement in American history. Anthropic agreed to pay $1.5 billion after downloading approximately 482,460 pirated books from Library Genesis and Pirate Library Mirror to train its Claude AI models.
If you’ve written a book, it may be on the list. Here’s what you need to know.
What Happened
In August 2025, Anthropic settled the class-action lawsuit Bartz v. Anthropic after a federal court issued a split ruling. Judge William Alsup determined that using copyrighted books to train AI models can constitute “transformative” fair use - but acquiring those books through pirate sites does not.
“You can’t just bless yourself by saying I have a research purpose and take any textbook you want,” Judge Alsup wrote.
The distinction matters. Anthropic didn’t just read copyrighted books - it downloaded over seven million of them from torrent sites. That’s infringement, regardless of what the company did with them afterward.
Facing potential statutory damages of $150,000 per willfully infringed work - potentially billions more than the settlement amount - Anthropic chose to settle.
Who Qualifies
Your book is covered if it meets these criteria:
- Has an ISBN or ASIN number
- Was registered with the U.S. Copyright Office within three months of publication, OR within five years of publication and before August 10, 2022
- Appears in the searchable database on the settlement website
The settlement covers approximately 500,000 titles. Search by author name, book title, ISBN, or publisher at anthropiccopyrightsettlement.com.
How Much Authors Receive
After legal fees (25%, or $375 million) and administrative costs, the settlement provides roughly $3,000 per covered work.
The default splits:
- Trade and university press books: 50% to authors, 50% to publishers
- Self-published or rights-reverted works: 100% to authors
- Multi-author works: Authors split their portion equally among co-authors
Parties can negotiate different arrangements with documentation. Authors who believe they deserve a larger share must provide evidence of their contract terms.
Key Deadlines
- March 30, 2026: Claims deadline. Miss this, and you waive your right to sue AND receive no payment.
- April 2026: Final fairness hearing
- June 2026: Payments begin
How to File a Claim
- Visit anthropiccopyrightsettlement.com
- Search the Works List for your books
- Complete the claim form with your payment details (ACH, Zelle, or check)
- Identify any co-authors or publishers
If you have many books, email [email protected] to work with the claims administrator.
The Bigger Picture
The Anthropic settlement establishes a precedent that’s rippling through the AI industry. The core principle: how AI companies acquire training data matters as much as what they do with it.
This creates a split legal standard. Courts have found that training on copyrighted works can qualify as fair use because AI models extract statistical patterns rather than reproducing creative expression. But obtaining those works through piracy triggers separate infringement liability.
The practical result: AI companies must now meticulously document their data sources or face massive litigation exposure.
Over 50 AI Copyright Cases Pending
Anthropic’s settlement comes as the AI industry faces an avalanche of copyright lawsuits:
OpenAI MDL (ongoing): The New York Times, Daily News, and Center for Investigative Reporting have consolidated cases in the Southern District of New York. A federal judge recently ordered OpenAI to produce 20 million ChatGPT conversation logs to plaintiffs.
Disney v. Midjourney (pending): Major studios allege unlawful copying for image generation. This case may establish standards for visual AI models.
Universal Music v. Anthropic (filed January 2026): Labels filed a $3.1 billion lawsuit alleging Anthropic built Claude “on a foundation of torrented piracy” - for music, this time.
Thomson Reuters v. Ross Intelligence (appealing): The Third Circuit is reviewing whether copying legal headnotes for AI training qualifies as fair use. The lower court said no.
What This Means for AI Training
The Anthropic settlement signals a shift toward formal licensing arrangements. While fair use may protect training on legitimately acquired content, the financial and reputational risks of using pirated materials are now impossible to ignore.
Companies like OpenAI have already begun negotiating licensing deals. In late 2025, Disney and OpenAI reached an agreement for access to the studio’s catalog - likely prefiguring how the industry will operate going forward.
For authors, the message is clear: check the settlement website before March 23. That deadline isn’t going to move.
The Bottom Line
If you’ve written a book with an ISBN, search the settlement database at anthropiccopyrightsettlement.com. You have two weeks to claim approximately $3,000 per title. After March 30, you lose both the payment and your right to sue.