Federal Judge Rules in Favor of Anthropic in AI Book-Training Lawsuit

Federal Judge William Alsup ruled that Anthropic’s use of published books to train its AI models without obtaining author permission is legal. This is the first court decision to support the argument that fair use can shield AI companies from liability when using copyrighted content to develop large language models (LLMs).
Image Credits:Lou Benoist / Contributor / Getty Images

Federal Judge William Alsup ruled that Anthropic’s use of published books to train its AI models without obtaining author permission is legal. This is the first court decision to support the argument that fair use can shield AI companies from liability when using copyrighted content to develop large language models (LLMs).

Alsup’s Ruling May Shape Future AI Copyright Battles, Tilting Courts Toward Tech Firms

The ruling is a setback for authors, artists, and publishers who have filed numerous lawsuits against companies like OpenAI, Meta, Midjourney, and Google. While Alsup’s decision doesn’t set a binding precedent for other courts, it could influence future cases and signals a potential judicial shift in favor of tech companies over content creators.

These lawsuits often hinge on how a judge interprets fair use—a complex and outdated exception in copyright law that hasn’t been significantly revised since 1976, long before the internet or the rise of generative AI.

Fair use decisions consider factors like the purpose of the use (e.g., parody or education), whether it’s for commercial benefit (you can write “Star Wars” fan fiction, but not sell it), and how much the new work transforms the original.

Companies like Meta have also defended their use of copyrighted material in AI training by invoking fair use, but until this week’s ruling, it was unclear how courts might respond.

Authors Allege Anthropic Built a Massive Book Database Using Pirated Copies

In Bartz v. Anthropic, the plaintiff authors also raised concerns about how Anthropic obtained and stored their books. The lawsuit alleges the company aimed to build a “central library” containing “all the books in the world” for permanent retention—and that millions of those books were downloaded illegally from pirate websites.

While the judge found that using the books for AI training qualified as fair use, the court will still hold a trial to examine the legality of how Anthropic built and maintains this “central library.”

We will proceed to trial over the pirated copies used to build Anthropic’s central library and any resulting damages,” Judge Alsup wrote in his ruling. “The fact that Anthropic later purchased a book it had previously downloaded illegally doesn’t erase the original act of infringement, though it may influence the amount of statutory damages awarded.”


Read the original article on: TechCrunch

Read more: Your Smartwatch Could Detect Illness Early and Aid Pandemic Prevention