Anthropic Wins Key Ruling in Copyright Lawsuit, But Faces Trial Over Pirated Books

A federal judge ruled in favor of Anthropic on Monday during a copyright case in the United States. District Judge William Alsup of San Francisco found that the AI company did not break the law by using millions of copyrighted books to train its chatbot Claude. However, the company must still face trial over pirated books stored.

According to the official ruling,  Anthropic purchased and downloaded millions of copyrighted books—many of them from pirate sites—for its “central library,” from which it uses various sets and subsets to train its large language models (LLMs).

Some of the authors whose works were included—Andrea Bartz, Charles Graeber, and Kirk Wallace Johnson—sued Anthropic for copyright infringement. Judge Alsup, however, determined that Anthropic made “fair use” of the collected material.

“Claude’s customers wanted Claude to write as accurately and as compellingly as Authors,” states the document, referring to the plaintiff writers as “Authors.” “So, it was best to train the LLMs underlying Claude on works just like the ones Authors had written, with well-curated facts, well-organized analyses, and captivating fictional narratives — above all with ‘good writing’ of the kind ‘an editor would approve of.’”

The judge also noted that the books used were part of the training process only. Claude’s public-facing version is controlled by software that filters outputs and prevents the generation of exact copies or traceable reproductions of the original texts.

“Like any reader aspiring to be a writer, Anthropic’s LLMs trained upon works not to race ahead and replicate or supplant them—but to turn a hard corner and create something different,” said Alsup.

The judge considered Claude’s use of the material “quintessentially transformative,” but raised concerns over the use of pirated copies. Anthropic reportedly downloaded over 7 million copies of books from pirate libraries. “Anthropic had no entitlement to use pirated copies for its central library,” stated the judge, adding that a separate trial will address this issue.

Just as other AI companies have been involved in legal proceedings—the BBC recently threatened Perplexity with legal action for scraping its content—Anthropic has already faced multiple cases related to the use of content created by creative professionals. A few weeks ago, a federal judge from California also ruled in favor of Anthropic in a music AI copyright lawsuit, and the AI company also took responsibility for AI hallucination in a copyright lawsuit.