The University of Tennessee Research Foundation has sued Anthropic, accusing the artificial intelligence company of using patented neural-network technology inside Claude Code without permission or compensation.
Filed on July 20 in the U.S. District Court for the District of Delaware, the lawsuit targets Anthropic’s coding assistant and its underlying agent-based software architecture, with the university foundation seeking damages, a permanent injunction and a jury trial.
Two patents at the centre of the fight
According to the complaint, Anthropic infringed U.S. Patent Nos. 10,019,470 and 10,095,718, which cover methods for constructing, adapting and reusing components within artificial neural networks.
The foundation said the patented work, developed years before generative AI became a mainstream commercial industry, addresses key challenges in neural-network design, optimization, monitoring and adaptive learning while enabling useful computational structures to be identified, reused and improved.
Researchers named in the filing include J. Douglas Birdwell, an emeritus professor whose work has focused on artificial intelligence and neuromorphic computing, and Mark Dean, the veteran computer engineer known for helping develop foundational IBM personal-computer technology.
Claude Code accused of using protected methods
The foundation alleges that Claude Code uses recurring background processes, memory-consolidation systems, multi-agent workflows and self-correcting feedback loops that fall within the scope of its patents.
Among the features highlighted are systems that reorganize stored information, monitor software activity, route tasks between AI agents and react to errors before they disrupt an external computer environment.
The complaint also points to Claude Code’s ability to interact with terminals, files, displays and other systems, arguing that these functions rely on configurable neural pathways and control structures described in the university’s patents.
Foundation seeks financial compensation and an injunction
The University of Tennessee Research Foundation said Anthropic directly infringed the patents by making, using and selling the accused technology. It also alleged indirect infringement, arguing that Anthropic encouraged customers to use Claude Code in ways that violated the patents.
The foundation is asking the court to award damages of at least a reasonable royalty, along with interest, legal costs and attorneys’ fees. It also wants Anthropic barred from continuing any infringement.
Anthropic’s court fight spreads across AI, music and data
Anthropic is also fighting Universal Music Group, Concord and ABKCO in a major copyright case alleging that Claude was trained on protected song lyrics and can reproduce them without permission, with the publishers arguing that the company cannot shield the alleged conduct under fair use.
Reddit is pursuing a separate lawsuit accusing Anthropic of scraping its platform more than 100,000 times and using user posts to develop Claude without a licensing agreement.
Chicken Soup for the Soul publisher has also sued Anthropic alongside Apple, Google, Meta, OpenAI, Nvidia, Perplexity and xAI, alleging that the companies used pirated copies of its books to train their AI systems
