Anyone else been following the #Artificial Intelligence vs. Copyright Law narrative very closely on the podcast these last few months? For those keeping track, Anthropic scored a major legal victory for Team “All Your IP Belong To Us” this week. US District Judge William Alsup ruled that using #The Book Club 📚 to train AI models is “transformative” enough to be considered “fair use.” The suit had been brought by a collective of three authors — Andrea Bartz, Charles Graeber, and Kirk Wallace Johnson — who argued that Anthropic built its business on top of a cache of “hundreds of thousands of copyrighted books.” But Judge Alsup writes that Anthropic’s AI training does not reproduce the entirety of any author’s work, and does not output results that are recognizably “one author’s identifiable expressive style.” He concludes, therefore, that an LLM ingesting a book for training is similar to “any reader aspiring to be a writer,” innocently and legally leafing through the pages themselves. This is a first-of-its-kind landmark ruling that could establish guardrails for AI training and IP law moving forward. In a statement, Anthropic said they were “pleased” with the outcome, and found Judge Alsup’s arguments consistent with their understanding of copyright law. It’s important to note, however, that Anthropic is not entirely out of the woods. Judge Alsup ordered a separate trial to consider whether the company used illegally obtained, “pirated” books to train its models and build out its digital library, and if so, to determine their potential damages and liability. https://www.cnbc.com/2025..