Federal Court Rules AI Training On Copyrighted Books Fair Use, Trial Scheduled Over Pirated Library

A federal judge in San Francisco ruled that Anthropic’s use of millions of copyrighted books to train its AI chatbot Claude qualifies as fair use under U.S. copyright law, marking a significant legal win for AI companies. Judge William Alsup emphasized that training AI models on copyrighted works is transformative, comparing it to a reader learning to write rather than copying or supplanting the originals. However, the court found that Anthropic’s downloading and maintaining a large library of pirated books violated copyright law, requiring a trial to determine the extent of infringement and potential damages. This ruling sets a precedent that could influence ongoing lawsuits against other AI firms like OpenAI and Google, which face similar copyright infringement claims. Despite the fair use victory for training purposes, Anthropic could still face substantial financial liabilities for the unauthorized use of pirated texts. The case highlights the complex balance between innovation in AI training and the protection of intellectual property rights.