A United States federal judge has ruled that Anthropic did not violate copyright laws by using books to train its Claude AI model, declaring the use “fair” and “transformative.”

In the decision, U.S. District, Judge William Alsup, said the Amazon-backed company’s large language models do not publicly reproduce protected creative content or replicate any author’s unique expressive style, according to CNBC.

“The purpose and character of using copyrighted works to train LLMs to generate new text was quintessentially transformative,” Alsup wrote. “Like any reader aspiring to be a writer.”

The ruling represents a key win for AI developers amid ongoing legal disputes over the use of copyrighted material in training large language models. Judge Alsup’s decision helps set early legal precedent, outlining both the limits and potential pathways for the industry.

A spokesperson for Anthropic said the company welcomed the ruling, describing it as “Consistent with copyright’s purpose in enabling creativity and fostering scientific progress.”

Filed in August in the U.S. District Court for the Northern District of California, the lawsuit was brought by authors Andrea Bartz, Charles Graeber, and Kirk Wallace Johnson. It claimed that Anthropic used hundreds of thousands of copyrighted books without permission to build a multibillion-dollar business.

A key part of the lawsuit focused on a collection of around 7 million books that Anthropic allegedly pirated and kept in a “central library.” While the materials were retained, the company ultimately chose not to use them to train its large language models.

Judge Alsup also ordered a trial to determine how the pirated books were used in creating Anthropic’s central library and to assess any potential damages resulting from their use.

“That Anthropic later bought a copy of a book it earlier stole off the internet will not absolve it of liability for the theft, but it may affect the extent of statutory damages,” the judge wrote.