Lawsuit Tests What Universities Are Owed for AI Research

Over a decade ago, when few outside tech circles and computer science departments could have imagined AI’s reach into everyday life, the University of Tennessee secured two patents for artificial neural networks. Last week the university’s research foundation sued Anthropic in a federal court in Delaware for infringing the rights to those patents. It’s the first time a university has sued an AI company and could open the doors to more universities claiming their stake in the AI revolution.

In its court filing, the University of Tennessee Research Foundation (UTRF) said the patented inventions “reflect years of research conducted at The University of Tennessee concerning biologically inspired approaches to machine intelligence.”

Tennessee is far from the only academic institution doing pioneering research that has led to discoveries now used in AI technology. Frontier AI companies—the likes of OpenAI, Anthropic and Google DeepMind—are built on knowledge developed in universities over the past century. The Turing test, a foundational measurement of machine intelligence, was published by British mathematician Alan Turing in 1950 while he was at the University of Manchester. Dartmouth College claims the term “artificial intelligence” was coined by its researchers in 1956. And researchers at Carnegie Mellon and UC San Diego, including the “Godfather of AI,” Geoffrey Hinton, developed the artificial neural networks that form the foundation for deep learning.

Anthropic’s own co-founder Dario Amodei got his doctorate degree from Princeton. His journey from academia to the tech world isn’t unusual in the mutually beneficial relationship that exists between higher education and industry.

Technology and knowledge transfer are hallmarks of this partnership. Research foundations like UTRF exist at institutions across the country to commercialize and license university-developed research so that it has the greatest real-world impact. Tens of thousands of patents have been filed by U.S. universities.

But until now, the boundaries of that transfer in the realm of AI haven’t been tested. Only textbook publishers and authors have attempted to claw back damages from AI companies, including Meta and Google through copyright-infringement suits. Others, such as Taylor & Francis Group and Wiley, have opted to enter licensing agreements with AI companies that will pay to train their models on the research produced in institutions.

What’s most notable about the Tennessee suit is that it takes issue with the architecture of the technology, not what is going into the machine for training. The university has laid claim to the years of scientific research and technological innovation that has enabled the creation of LLMs and the multibillion-dollar companies that are developing them. The suit demands recognition not for what’s making the algorithms smarter, but for the university’s part in making them possible.

UTRF’s filing didn’t mention the exact amount it’s seeking in damages, just compensation “adequate for the infringement” and “no less than a reasonable royalty.” Anthropic’s market value of $1.2 trillion promises a handsome claim if the challenge is successful. For context, Carnegie Mellon sued semiconductor manufacturer Marvell Technology Group Ltd. for patent infringement in 2016 and settled for $750 million. While the federal government is cutting, limiting and attempting to politicize research funding, higher ed could benefit from protecting its innovations more aggressively.

If past patent-infringement cases are anything to go on, the suit could drag on for years and be unsuccessful. After an expensive, decade-long legal battle, in 2024 the University of Wisconsin lost its own case against Apple that argued chips used in iPhones and iPads violated its patent rights. But beyond the potential payouts of this case and any that follow, simply by making the claim, universities and researchers stand to gain recognition for their role in advancing AI technology. And, maybe, more influence in how AI is developed in the future.