Anthropic, the AI company behind Claude, has reached a $1.5 billion class action settlement resolving allegations that it used copyrighted books without authorization to train its large language models. This is one of the largest AI copyright settlements in history.
If you're an author, publisher, or rights holder of written works published before 2024, you may be owed significant compensation — potentially $3,000 per qualifying work.
What Are the Allegations?
The lawsuit alleges Anthropic scraped millions of books, articles, and written works from the internet — including pirated sources like Library Genesis — and fed them into Claude's training data without obtaining licenses or paying rights holders. Plaintiffs include prominent fiction authors, journalists, and publishing houses.
Anthropic settled without admitting liability, stating it used "publicly available data" and maintains that training AI on text constitutes fair use. Courts have not resolved the underlying legal question — the settlement sidesteps that fight.
Who's Eligible?
You may file a claim if:
- You own copyright to books, articles, blog posts, or other written works published before January 1, 2024
- Your work was published in English and publicly accessible (physical, digital, or online)
- You're a US rights holder (foreign rights holders may qualify under some sub-classes)
Self-published authors qualify alongside traditionally published authors. Academic papers, textbook authors, and long-form journalists may also be eligible.
How Much Can You Claim?
The settlement structure pays:
- Up to $3,000 per qualifying work for registered copyright holders with strong evidence
- Base payment tiers for works without formal registration, scaled by length and publication date
- Publisher class members may claim based on their catalog size
Final payouts depend on total validated claims — with a $1.5 billion fund and relatively limited awareness among authors, payments could be substantial.
How to File
- Visit the Anthropic AI Copyright settlement claims portal
- List your qualifying works (title, publication year, ISBN or URL)
- Upload copyright registration certificates if available (not required but increases payment)
- Certify your ownership of rights
- Submit before the February 5, 2027 deadline
The Larger Trend: AI Training Data Settlements
This is not the first AI copyright settlement and won't be the last. OpenAI faces similar litigation from the New York Times and multiple author coalitions. Getty Images sued Stability AI over image training data. The Anthropic settlement establishes a damages framework that other pending cases will likely reference.
For creators, the message is clear: the legal system is catching up with AI training practices. Settlements like this one may become a recurring income stream for prolific authors over the next decade as more AI companies face accountability for their training datasets.
Act Before February 2027
With a deadline over a year away, it's easy to procrastinate. Don't. The claims process takes 15–30 minutes for a detailed submission and could return thousands of dollars per book. Track this settlement on SettlementRadar and set a reminder.
