A U.S. federal judge has ruled that AI company Anthropic did not break copyright laws by training its Claude AI model on books that were legally purchased and scanned, calling such use “quintessentially transformative” and protected under fair use. However, the court made a clear distinction regarding the use of pirated books, stating that if Anthropic used illegally downloaded materials from online shadow libraries, it would not fall under fair use. That part of the case will go to trial in December to determine responsibility and potential damages. This ruling sets a significant precedent for the AI industry, affirming that training on lawfully obtained content is permissible, while also warning companies of serious legal consequences for using pirated data. The outcome could impact ongoing and future lawsuits involving other major AI developers like OpenAI, Meta, and Google.
4
