James Grimmelmann is a professor of law at Cornell University and a leading expert on copyright law. Grimmelmann walks through the complex process courts use to determine whether training AI models on copyrighted materials—like OpenAI using New York Times articles—is infringement or fair use. He highlights key precedents like the Google Books case, emphasizing how courts weigh transformative uses against potential market harms.

The discussion addresses the nuances of generative AI, notably cases where models inadvertently reproduce large excerpts from training materials. Grimmelmann argues that while the industry has largely addressed explicit "regurgitation," ambiguity remains around subtler forms of copying, particularly with image-generating models, which could substantially impact copyright holders like Getty Images.

Grimmelmann and the hosts delve into potential legal outcomes, including moderate rulings that force licensing agreements, or harsher ones that could significantly restrict the availability of open-source AI models. The interview also touches on Congress's historical reluctance to intervene in contentious digital copyright issues, leaving critical decisions to be gradually shaped by court rulings.

Dean and Tim conclude that while an outright shutdown of generative AI by courts is improbable, the forthcoming legal decisions will likely reshape the industry's structure, potentially favoring larger companies capable of negotiating extensive licensing deals. Grimmelmann anticipates initial district court rulings within the year and appellate decisions by 2026, setting the stage for a pivotal shift in how AI companies use copyrighted works.

This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit www.aisummer.org