Skip to main content

Dale Nelson Authors Forbes Article Unpacking Major AI Fair Use Rulings

Dale Nelson recently penned a Forbes piece delving into two recent rulings in lawsuits involving Meta and Anthropic, respectively, that deemed the use of copyrighted books for training AI models to be fair use. The article, titled “Copyrighted Books Are Fair Use For AI Training. Here’s What To Know,” explains that although these lawsuits involved very particular facts, both rulings lay out guidelines as to how the decision might affect other cases. Dale writes that the Meta and Anthropic cases can shed light on how future AI systems can be designed not to infringe on copyright. 

Dale unpacks the ruling in Bartz v. Anthropic, where a group of three authors sued the AI firm after it utilized their works to train Claude, its chatbot. Dale writes that the judge determined the use of the books to train Claude was “quintessentially transformative,” and that Anthropic’s actions in purchasing print books and converting them to digital “was merely format-shifting for space and search capability purposes,” and, as such, was held to fall under fair use.

Two days later, in the Kadrey v. Meta ruling, another judge ruled that Meta was justified in using the works of thirteen authors to train its generative AI model, Llama. However, Dale highlights that the judge “did so very reluctantly, chiding the plaintiff’s lawyers for making the ‘wrong’ arguments and failing to develop an adequate record.” She goes on to add that the judge clearly emphasizes his fears of “AI systems potentially flooding the market with substitutes for human authorship and destroying incentives to create” in his decision.

Dale concludes by outlining several of her takeaways about a path to non-infringing AI systems based on the outcomes of both cases. She explains that using copyrighted books and creating a central library of books for AI training will fall under fair use. However, she notes that “the system must be designed so that the output of your system does not reproduce any source material books in an exact or substantially similar way.”

Read the full article in Forbes.