Anthropic Settled Authors' Lawsuit for $1.5 Billion
The 2025 court ruling follows unauthorized AI training on digital library collections.
Updated on Sept. 30, 2026 in Artificial Intelligence
Live Poll
Should AI companies be required to pay creators for using their work to train models?
Anthropic agreed to a $1.5 billion settlement for authors and publishers after a 2025 class action lawsuit regarding unauthorized book downloads. The settlement compensates rights holders for 500,000 titles used to improve artificial intelligence models.
Why it matters
The ruling penalizes the unauthorized scraping of digital libraries for AI training data, establishing financial liabilities for the practice. Publishers and authors are now verifying ownership rights to claim shares of the total settlement pool.
Anthropic reached the $1.5 billion settlement to cover 500,000 eligible titles, with an estimated payout of $3,000 per book. This total accounts for only a fraction of the 7 million copies originally scraped from repositories like LibGen and PiLiMi.
The players
Anthropic
An artificial intelligence lab focused on developing large language models and safety-aligned generative AI systems.
William Alsup
A United States District Judge for the Northern District of California who presided over the class action.
Authors Guild
A professional organization that advocates for writers' rights and manages the settlement eligibility database.
The details
The lawsuit centered on the mass ingestion of data from digital repositories LibGen and PiLiMi, which Anthropic used to refine its large language models—algorithms trained on massive datasets to predict and generate text. Judge William Alsup of the U.S. District Court for the Northern District of California ruled that the unauthorized downloading and storage of these copyrighted works constituted an infringement. To claim funds, authors must verify their rights through the Authors Guild portal and submit required reversion-of-rights documentation to an administrator.
Timeline
2025: Judge Alsup ruled on the class action case.
Mid-November 2026: The first portion of the settlement payment is expected.
2027: The second portion of the settlement payment is expected.
The Tech Race
This settlement sets a high-water mark for litigation involving the unauthorized ingestion of copyrighted data for model training. It marks a departure from earlier research-stage disputes, forcing labs to account for intellectual property in their compute-intensive scaling strategies.
Authors can check their eligibility through the official Authors Guild settlement eligibility database to determine if their titles qualify for the $3,000 per-title payout. Meanwhile, users may notice increased friction in content distribution, such as Amazon's new two-book-per-week limit on self-published uploads.
The takeaway
This case establishes that mass-scraping libraries for training data carries massive financial liability. Affected authors should continue monitoring the administrator's portal for claim processing updates leading up to the first payments in November 2026.
Further reading
For more on how companies are handling data acquisition, visit Artificial Intelligence.
More information
Authors should verify their rights via the Authors Guild settlement eligibility database.
Source note: This article includes information reported by Concord Monitor.
Live Poll
Should AI companies be required to pay creators for using their work to train models?
