Can an AI Platform be Listed as an Inventor on a Patent?
This blog post addresses a frequently asked question – can an AI platform (e.g. large language model (LLM)) be a named inventor on a patent? The short answer is no. However, the circumstances involving when humans can be named inventors who use AI platforms will continue to evolve. There is significant nuance to this determination as demonstrated by the recent USPTO guidelines which will be discussed here.
While an LLM may be used to propose ideas and concepts when prompted by humans, LLMs cannot be named as inventors. As of the time of this post, it is well established that patents cannot name AI platforms as inventors, only humans.[1] Inventorship should be predicated upon human inventive contributions, including inventions inspired by using AI platforms as a tool during the conception process.[2]
For example, if the inventor of the wheel had access to AI platforms, she might have written prompts asking about the best materials to use for the wheel. The AI platform could return a list of options of materials with pros and cons for each. The patent’s detailed specification could disclose various embodiments of the wheel with the different materials including a discussion of their pros and cons. The patent could conclude with claims broadly directed to the wheel and narrower claims that include the wheel with the various materials. In short, the human would be the named inventor of all these claims.[3] This example is consistent with the recent USPTO guidelines posted on February 13, 2024.[4]
Even though a human may use an AI platform in the process of conceiving an invention, use of the AI platform does not negate the human’s contributions as an inventor, so long as the human contributes significantly to the invention. The contributions necessary to constitute “significant contributions” will be determined on a case-by-case basis. The guidance provides that “a significant contribution could be shown by the way the human constructs the prompt in view of a specific problem to elicit a particular solution from the AI system.”[5]
However, with respect to the example, let’s assume that the AI platform not only suggests different materials to use for the wheel, but doing away with having wheels entirely and using a magnetic levitation device instead. Who is the inventor of the magnetic levitation device, the human or the AI platform? The answer is neither. Since a human did not contribute to the conception of the magnetic levitation device, the human cannot be the inventor and since the AI platform is not human, it cannot be named as an inventor either.
This will likely be an evolving area of the law. Future blog posts will discuss future rulings from the USPTO on this subject, judicial pronouncements that further refine or counter the guidance from the USPTO, and/or changes in the statutory requirements under 35 USC 115 & 116 concerning inventorship.
Bob Steinberg is the Founder of Generative IQ LLC, a venture fund that provides capital to early-stage IP rich AI start-ups. He has been protecting and litigating IP rights, working with technology entities and entrepreneurs navigating the IP landscape and monetizing blocking rights for over 30 years.
[1] Thaler v. Vidal 43 F.4th 1207 (Fed. Cir. 2022).
[2] USPTO October 2020 Report on AI.
[3] Biden AI Executive Order (EO) on AI (Oct. 30th, 2023) (EO requires the USPTO to issue guidance concerning various topics, including how use of AI generated content during the inventive process may affect inventorship. For example, guidance on when human inventors can use AI to help conceive or reduce to practice the invention and when AI infrastructure is eligible for patents).
[4] USPTO Inventorship Guidance Federal Register Feb. 13, 2024 (this guidance applies to all applications, and to all patents resulting from applications, filed before, on or after February 13, 2024).
[5] See note #4, USPTO Inventorship Guidance Federal Register Feb. 13, 2024.
