On November 26, 2025, the United States Patent and Trademark Office (USPTO) issued revised guidelines for determining inventorship of AI-assisted inventions,[1] which is a significant development in life science innovations, which increasingly rely on computational and AI-driven tools. The USPTO rescinded its earlier guidance[2], which suggested applying the Pannu factors[3] for determining joint inventorship to the determination of whether a single person using an AI system is an inventor. The new guidance emphasizes that the touchstone of inventorship remains “conception,” the formation in the mind of the inventor of a definite and permanent idea of the complete and operative invention.
For life science innovations, the new guidance clarifies that this traditional legal standard of conception applies regardless of whether AI is used. Therefore, when a single person is involved in the inventive process, the only question is whether that person conceived the invention under the conception standard. The USPTO characterizes AI systems as analogous to laboratory tools, such as next-generation sequencing platforms, computational chemistry suites, laboratory automation systems, or high-content screening systems, that assist in research and development but do not themselves conceive inventions. For example, AI systems may generate candidate molecules, predict biomarkers, suggest gene edits, or optimize cell-culture conditions, but they remain tools used by the human scientists so that there remains the potential for conception by the humans using them.
The Pannu factors continue to govern situations with multiple human contributors, including in multidisciplinary technologies such as are common in the life sciences, where computational biologists, formulation scientists, protein engineers, and researchers collaborate while using AI-assisted platforms.
Although U.S. courts have affirmed that only human persons can be inventors[4], they have not yet addressed the boundaries of human inventorship when AI shapes experimental design or otherwise helps a person conceive an invention. Therefore, to reduce prosecution risk and maintain patent enforceability, biotech companies using AI tools for various applications, such as drug discovery and design, analyzing complex data sets, improving manufacturing efficiency, and developing biologics should consider maintaining detailed concurrent documentation of the inventive process, including conception, that ultimately lead to the claimed invention to assist in demonstrating human inventorship. Companies may want to document how human scientists are engaged in the development, improvement, and validation of AI-assisted processes. Examples of records that may assist in demonstrating human inventorship include experiment designs, prompt engineering, data analysis choices, modifications, and iterations. As other countries may develop different legal standards for inventorship of AI-derived inventions, it will also be important to be aware of how disclosures could affect inventorship claims for global patent prosecution.
Editor: Brenden S. Gingrich,Ph.D.
[1] https://public-inspection.federalregister.gov/2025-21457.pdf?utm_campaign=subscriptioncenter&utm_content=&utm_medium=email&utm_name=&utm_source=govdelivery&utm_term=
[2] https://public-inspection.federalregister.gov/2024-02623.pdf
[3] An inventor must (1) contribute in some significant manner to the conception or reduction to practice of an invention, (2) make a contribution to the claimed invention that is not insignificant in quality when measured against the full invention, and (3) do more than merely explain well-known concepts and/or the current state of the art.
[4] https://www.cafc.uscourts.gov/opinions-orders/21-2347.OPINION.8-5-2022_1988142.pdf
Subscribe for News and Updates
