In a ruling that could have wide-ranging implications for the artificial intelligence (AI) industry, a federal judge in San Francisco has reportedly decided that Anthropic’s use of copyrighted books to train its AI model Claude falls under “fair use”. However, the same court also found that Anthropic had infringed copyright by storing pirated copies of books in a digital library, according to a report by Reuters. The decision, handed down on Monday by US District Judge William Alsup, is seen as a partial victory for the AI firm. The judge concluded that Anthropic’s use of the authors’ books for training its AI was legal, describing the use as “exceedingly transformative” and in line with the principles of fair use outlined in US copyright law.

“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 Judge Alsup.

Anthropic, which is backed by Amazon and Google, was sued last year by authors Andrea Bartz, Charles Graeber, and Kirk Wallace Johnson. The writers argued that Anthropic used pirated versions of their books without permission or compensation to teach its Claude model how to respond to user prompts. The lawsuit is one of several filed against major AI developers, including OpenAI, Meta, and Microsoft, over copyright concerns.

An Anthropic spokesperson welcomed the court’s decision on fair use. “We are pleased that the court recognised our AI training was transformative and consistent with copyright’s purpose in enabling creativity and fostering scientific progress,” the spokesperson said.

However, Judge Alsup also ruled that Anthropic crossed a legal line by downloading and storing over seven million pirated books in what was described as a “central library of all the books in the world.” This storage was not protected by fair use and amounted to copyright infringement, the judge said. A trial is scheduled for December which will determine how much Anthropic owes the authors for the violation.

Under US law, willful copyright infringement can result in damages of up to $150,000 per work. The court will examine the scale and impact of the infringement before deciding on the amount.

Judge Alsup dismissed Anthropic’s argument that the source of the books was irrelevant to fair use. “This order doubts that any accused infringer could ever meet its burden of explaining why downloading source copies from pirate sites that it could have purchased or otherwise accessed lawfully was itself reasonably necessary to any subsequent fair use,” he wrote.

The ruling comes as generative AI companies face increased legal scrutiny over how they acquire and use copyrighted material. Fair use has become a key defence for tech firms, who argue that their AI systems produce new and transformative content rather than mere copies.

Anthropic had argued in court that training AI on existing content is not only allowed but encouraged by copyright law because it promotes innovation. The company said it used books to study writing styles, extract ideas that aren’t protected by copyright, and build advanced technology.

Authors, however, argue that this use threatens their ability to make a living. They say that training AI models on their work without permission creates tools that can generate competing content, putting their careers at risk.

While the judge’s fair use ruling favours Anthropic, the finding of copyright infringement due to storage practices underscores the legal grey area in which AI companies operate. With other lawsuits ongoing, this case could set a precedent for how courts interpret the use of copyrighted content in AI development.

The ruling also follows a recent lawsuit filed by Reddit against Anthropic for allegedly scraping user comments without consent to train Claude. While that suit focuses on breach of contract rather than copyright, it adds to the growing pressure on AI companies to respect digital content boundaries.

Anthropic, founded in 2021 by former OpenAI executives, is considered one of the leading challengers to OpenAI. With backing from Amazon and Google parent Alphabet, its Claude chatbot has been integrated into products such as Alexa.

– Ends