Can AI Use Copyrighted Books? US Judge Says Yes

In a significant legal decision, U.S. District Judge William Alsup ruled that Anthropic, an artificial intelligence company behind the chatbot Claude, did not violate copyright laws when it trained its model using published literary works without author consent.

The judge declared the practice “fair use,” emphasizing that Claude’s outputs were transformative and didn’t merely copy existing books.

Anthropic argued that, much like a student learning from reading, its model absorbs information to generate new ideas, not replicate old ones.

However, the ruling wasn’t entirely in the company’s favor. Alsup also determined that Anthropic’s storage of over 7 million pirated books in a centralized library system did infringe on copyright and will proceed to trial in December 2025.

The case reflects broader tensions between creators and AI developers as artificial intelligence continues to absorb vast amounts of publicly available content. With tech companies defending their practices under “fair use” and authors demanding stronger protections, the verdict may shape future regulations.