With the rise of generative AI, there seems to be an issue of blurring boundaries when it comes to original and reproduced content, as well as the economic threat it tends to pose to creative professionals. This controversy is especially applicable when it comes to training AI models like ChatGPT or Claude, and where massive datasets are used to help the models learn and be able to generate new content. Anthropic has been under hot water for quite some time, and has been accused of training its Claude AI models using copyrighted material. However, the case seems to be finally shaping up in favor of Anthropic as a judge recently ruled that the method of AI training was fair use under U.S. copyright law.
Anthropic wins in the AI copyright case as a judge recently ruled that the training method used for its AI models falls under fair use
Anthropic has been fighting on the legal end and facing accusations of using copyrighted books to train its Claude AI model. U.S. federal judge William Alsup of the Northern District of California, however, passed a major ruling on June 24, 2025, in favor of Anthropic, where it was explicitly stated that the use of legally purchased and digitized books to train the Claude AI model is deemed fair use under the U.S. copyright law. The judge highlighted how turning text into AI knowledge and not copying or redistributing does meet the criteria of what is classified as fair use.
While Alsup ruled out that using legally obtained copyrighted content for training generative AI models and to learn is not copying and marked the clear distinction between the two, the judge, however, held Anthropic accountable for using pirated books from sites such as Book3 and LibGen and showed little tolerance for illegal data sourcing even if the intention behind it is transformative.
Judge Alsup further maintained that a separate trial would be held for Anthropic's use of pirated content and to decide the damages it would have to pay. Alsup maintained:
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.
By having the case turn two-fold, the court left the door open for authors to pursue a separate piracy trial of Anthropic. This is a major turning point for the future course of AI and its way of training AI models. By ruling that while training on legally purchased books falls under fair use, acquiring data through piracy cannot be justified, the judge is setting a strong precedent for tech companies and any AI cases going forward.
Follow Wccftech on Google to get more of our news coverage in your feeds.
