R&D Recordkeeping Is Best Remedy for New AI Invention Guidance

The US Patent and Trademark Office has grappled publicly with artificial intelligence’s role in the invention process since at least August 2019, when it first solicited public comments on patenting AI-assisted inventions. But the USPTO’s new official guidance released Feb. 13 seems to plow little new ground.

The guidance clarifies that no invention can be patented unless it was conceived by a human, which is consistent with the USPTO’s 2020 decisions rejecting patent applications naming only the DABUS AI system as the inventor. It also confirms that human inventors may use AI tools, but the result isn’t patentable unless humans make a significant contribution to the conception of the claimed invention.

So what should patent applicants do now? To start, they should review their research and development documentation protocols to ensure they’ll have the evidence in hand to support the application of the principles laid out in the guidance.

Thorough and consistent documentation at every stage of R&D—together with documentation of meetings, decisions, and feedback, including the prompts used with generative AI tools—will be invaluable if questions arise over whether humans made a significant contribution to the conception of an AI-assisted invention.

Patent applicants also should remember to regularly review and update documentation as each project progresses.

Non-Exhaustive Principles

The guidance includes five non-exhaustive principles to aid applicants and patent examiners in determining whether a given invention that resulted from the use of AI may nevertheless be patented.

The exact wording of the principles likely will be significant in responding to office actions and eventual appeals to the Patent Trial and Appeal Board and beyond. Patent owners can expect these principles to come up when an opponent challenges a patent’s validity:

• The use of an AI system in the process doesn’t by itself disqualify the invention from being patented.
• A significant contribution could be shown by the way the person constructs the prompt in view of a specific problem to elicit a particular solution from the AI system.
• A person who takes the output of an AI system and makes a significant contribution to the output to create an invention, or who conducts a successful experiment using the AI system’s output, may be a proper inventor.
• A natural person who designs, builds, or trains an AI system in view of a specific problem to elicit a particular solution could be an inventor, where the designing, building, or training of the AI system is a significant contribution to the invention created with the AI system.
• A person simply owning or overseeing an AI system used in the creation of an invention, without providing a significant contribution to the conception of the invention, doesn’t make that person an inventor.

The USPTO has provided two examples of how to apply these principles in actual invention scenarios—one involving a toy car, and the other involving drug discovery.

Work in Progress

While these principles will be applied to all pending and new patent applications as of Feb. 13, many details remain to be resolved as the USPTO applies this guidance in real time.

For businesses that engage in R&D leading to patenting of their inventions in the US, at the very least, this guidance adds important record-keeping requirements if any form of AI is used in the research process, and especially if used in conceiving of the invention.

The USPTO also is accepting public comments through May 13 to further refine the guidance.

This article does not necessarily reflect the opinion of Bloomberg Industry Group, Inc., the publisher of Bloomberg Law and Bloomberg Tax, or its owners.

Author Information

Rachel Reid is partner and head of artificial intelligence, US, at Eversheds Sutherland.

Ann Fort is partner at Eversheds Sutherland with focus on intellectual property rights.

Write for Us: Author Guidelines