Judge Approves Anthropic's Record $1.5 Billion Author Settlement

A U.S. federal judge has given final approval to Anthropic’s $1.5 billion class-action settlement with authors and publishers over pirated books used to train its Claude AI chatbot. The case, Bartz v. Anthropic, marks the largest documented copyright settlement in U.S. history.

Judge William Alsup of the U.S. District Court for the Northern District of California issued the approval after preliminary clearance on September 25, 2025 [7]. Anthropic reached the agreement in late August 2025 on the eve of a scheduled December 1 trial [2].

The settlement resolves claims that Anthropic downloaded roughly 7 million books from piracy sites Library Genesis and Pirate Library Mirror without permission [2]. About 482,000 to 500,000 registered works with U.S. Copyright Office ISBN or ASIN numbers qualify for the class [3]. Each eligible rightsholder stands to receive approximately $3,000 to $3,100 per work [4].

Anthropic will pay the $1.5 billion plus interest in four installments. The company already transferred the first $300 million after preliminary approval and must pay another $300 million within five days of final approval [7]. Plaintiffs’ lawyers seek $300 million in attorney fees, or 20 percent of the total [6].

The settlement requires Anthropic to destroy the original pirated files within 30 days and certify that the datasets were used in any commercially released models [7]. It covers only past legacy uses and grants no future license for the works [3].

Judge Alsup’s earlier June 2025 summary judgment ruled that training AI on lawfully acquired books can qualify as fair use, but using pirated copies does not [2]. The case began in 2024 when authors Andrea Bartz, Charles Graeber, and Kirk Wallace Johnson sued over unauthorized use of their books [1].

The ruling arrives as multiple copyright lawsuits against AI companies remain active. Earlier mass-digitization cases like Google Books established some fair-use precedents for large-scale text use, but none involved explicit piracy allegations [5].

For African and Nigerian AI startups, the decision clarifies that data acquisition methods carry significant legal risk. Companies scraping web content or academic libraries for training must now document provenance to avoid similar liability when serving global markets. The requirement to destroy pirated files and maintain auditable datasets offers a compliance template other jurisdictions may adopt.

African authors and publishers gain a precedent for collective action. Rights organizations may pursue comparable strategies if AI firms train on regional literary or educational content without licenses. Regulators at Nigeria’s NITDA and African Union bodies are likely to reference the fair-use versus piracy distinction when drafting AI copyright guidance.

This groundbreaking settlement will represent the largest publicly documented copyright recovery ever.

The settlement leaves open questions about ongoing suits against other AI developers. Observers will watch how companies adjust training data practices and whether additional class actions emerge from creative industries.