AI’s intellectual-property reckoning will not end with training data. As AI products evolve from passive models into persistent agents that remember, schedule, coordinate, decide and act...

2026-07-23 · Source: Pascal’s Substack · Field: Legal & Regulatory — Intellectual Property & Patents, Compliance & Risk Management, Artificial Intelligence & Machine Learning · Depth: Advanced, long

Summary

The University of Tennessee Research Foundation (UTRF) has filed a lawsuit against Anthropic, alleging that its Claude Code and agentic architecture infringe US Patent Nos. 10,019,470 and 10,095,718. These patents originate from neuroscience-inspired computing research. UTRF seeks damages and a permanent injunction, arguing that Claude's software features, such as schedulers, memory systems, and agent components, are equivalent to patented neuromorphic network elements like neurons, synapses, and central pattern generators. While the complaint is technically detailed and may proceed to discovery, its core infringement theories are vulnerable due to reliance on metaphorical rather than architectural equivalence. This case opens a new front in AI intellectual property litigation beyond training data, highlighting the need for AI developers to conduct patent freedom-to-operate reviews and carefully manage technical communications.

Key takeaway

For AI Architects and Directors of AI/ML developing agentic systems, you must proactively integrate patent freedom-to-operate reviews into your product lifecycle. This lawsuit demonstrates that even metaphorical language in public documentation can be used as evidence. You should conduct detailed claim-by-claim architecture analyses, document design-around decisions, and establish a licensing strategy to mitigate risks from neuromorphic and control-system patents before product launch.

Key insights

AI patent litigation is shifting from data to architecture, challenging software features as neuromorphic hardware equivalents.

Principles

• Patent infringement doesn't require copying.
• Technical metaphors are not legal proof.
• Public documentation can become infringement evidence.

Method

Conduct patent freedom-to-operate reviews, design around difficult claims, explore licensing early, and improve technical communication governance to distinguish metaphors from architecture.

In practice

• Map patent claims to product architecture.
• Separate LLM from UI rendering systems.
• Document design-around decisions.

Topics

• AI Patent Litigation
• Neuromorphic Computing
• Agentic AI Systems
• Intellectual Property Law
• Patent Freedom-to-Operate
• Claude Code

Best for: CTO, VP of Engineering/Data, Executive, Legal Professional, AI Architect, Director of AI/ML

Related on AIssential

• Intellectual Property vs. Machine Learning: Data Ownership and AI Models in the Age of Litigation — AI's reliance on vast datasets without clear consent creates profound, unresolved intellectual property and ownership disputes.
• What A Time For Legal AI! + Legal Innovators California — The legal AI market is rapidly expanding with new agentic AI solutions enhancing drafting, workflow, and contract management.
• Apple takes OpenAI to court — Apple's lawsuit against OpenAI highlights the intense competition and intellectual property risks in the rapidly evolving AI hardware sector.
• Microsoft built supercomputer to help OpenAI infringe copyrights, NYT alleged — New legal precedent requires plaintiffs to prove intentional inducement for contributory copyright infringement in AI training cases.
• AI Weekly Issue #484: Your AI chats can be used against you in court — AI's rapid, unpredictable evolution is challenging established legal, financial, and social frameworks.
• CNN says Perplexity is not merely learning from CNN in some abstract model-training sense, but repeatedly copying CNN content, retrieving it in real time,... — AI copyright litigation is shifting from training data to direct content reproduction and market substitution by "answer engines."

See Counsel's argued verdicts on the open AI decisions leaders are weighing →

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Pascal’s Substack.