In a critical win for the AI industry, a federal judge in San Francisco ruled that AI model training on books constitutes fair use under US copyright law. This is the first ruling of its kind and one that every stakeholder has been waiting and watching for.
The lawsuit in question, brought by three authors, is against Anthropic (the maker of Claude). The judge’s ruling says, “Authors contend generically that training LLMs will result in an explosion of works competing with their works. … This order assumes that is so. But Authors’ complaint is no different than it would be if they complained that training schoolchildren to write well would result in an explosion of competing works. This is not the kind of competitive or creative displacement that concerns the Copyright Act.”
In this specific case, the plaintiffs did not try to argue that AI models output infringing works. Instead, the case was strictly focused on inputs. The judge notes, “Authors concede that training LLMs did not result in any exact copies nor even infringing knockoffs of their works being provided to the public. If that were not so, this would be a different case. Authors remain free to bring that case in the future should such facts develop.” That seems a likely and important factor in other cases, especially those related to photography, artwork, and things like Disney characters. (Disney just filed a lawsuit against Midjourney.)
Ultimately, the judge finds training LLM models on authors’ books to be transformative and “spectacularly so,” a key factor for fair (legal) use. Another interesting facet of the ruling: The judge finds it fair use for Anthropic to digitize print books for the purposes of training if such copies are legally acquired.
However—in a significant win for authors and rights holders—the judge finds that Anthropic’s use of pirated works to train its models constitutes copyright infringement. A decision will be made in December on damages to be paid, and the damages may be astronomical if the case is class certified. The lowest statutory damage for this type of copyright infringement is $750 per book; Anthropic’s library of pirated titles consisted of at least 7 million books.
The Authors Alliance has a good starting analysis of the ruling.
Other lawsuits of a similar nature are underway in different courts; other judges could rule differently. In the July 2 issue, I’ll have an interview with an expert to deconstruct the implications of the ruling and what might happen next. I have to wonder if this will be motivation for plaintiffs and defendants alike to settle out of court.
Jane Friedman has spent her entire career working in the publishing industry, with a focus on business reporting and author education. Established in 2015, her newsletter The Bottom Line provides nuanced market intelligence to thousands of authors and industry professionals; in 2023, she was named Publishing Commentator of the Year by Digital Book World.
Jane’s expertise regularly features in major media outlets such as The New York Times, The Atlantic, NPR, The Today Show, Wired, The Guardian, Fox News, and BBC. Her book, The Business of Being a Writer, Second Edition (The University of Chicago Press), is used as a classroom text by many writing and publishing degree programs. She reaches thousands through speaking engagements and workshops at diverse venues worldwide, including NYU’s Advanced Publishing Institute, Frankfurt Book Fair, and numerous MFA programs.
