A federal judge approved a $1.5 billion settlement between Anthropic and major book publishers over allegations the AI company used pirated copyrighted material to train its Claude language model.

The settlement resolves a class-action lawsuit filed by authors and publishers who claimed Anthropic scraped millions of books without authorization to develop Claude's conversational abilities. The case was one of several high-profile copyright disputes targeting AI companies over their training data practices.

Under the agreement, Anthropic will pay $1.5 billion over three years to a fund that will compensate rights holders whose works were allegedly used in training. The company also agreed to implement new licensing procedures for future model development.

Publishers claim victory in landmark case

The Authors Guild and Association of American Publishers, who led the lawsuit, called the settlement a watershed moment for creator rights in the AI era. They argued that AI companies had systematically infringed copyright by using books, articles, and other content without permission or compensation.

Anthopic did not admit wrongdoing as part of the settlement but acknowledged the need for "sustainable partnerships with content creators," according to court filings. The company said it would work with publishers to establish licensing frameworks for future training datasets.

The settlement comes as AI companies face mounting legal pressure over their training practices. OpenAI and other major labs are defending similar copyright lawsuits from news organizations, artists, and writers.

Industry observers said the Anthropic settlement could set a precedent for how AI companies compensate content creators, potentially reshaping the economics of model training. The agreement requires court approval of the compensation distribution plan within 90 days.

Anthopic's Claude models compete directly with OpenAI's GPT series and other leading conversational AI systems. The company has raised over $7 billion from investors including Google and Amazon to develop what it calls "constitutional AI" focused on safety and alignment.