Overview
- The California federal court formally approved the roughly $1.5 billion settlement on July 20, 2026 to resolve a 2024 group lawsuit by authors who accused Anthropic of training its generative AI on copyrighted books without permission.
- The court drew a clear legal line by ruling that digitizing and using books legally purchased is lawful but that using about 7 million books taken from pirate websites amounts to copyright infringement.
- The parties had agreed on a settlement proposal in September 2025 and the judge’s July approval makes the payment binding while leaving the court’s source-based finding as legal guidance for future cases.
- Industry effects could include tighter data-acquisition controls, more licensing deals with publishers and authors, and increased compliance checks by AI developers to avoid using pirated material.
- Authors and rights holders stand to receive the settlement funds, and the decision gives them a practical remedy and leverage in negotiations over how copyrighted works are used to train large language models.