USPTO Issues New Guidance on Human Inventorship in AI-Assisted Patent Applications

Two recent developments, one in the United States and another in the United Kingdom, have provided greater clarity on the intersection of patent law and artificial intelligence (AI), particularly concerning the patentability of AI-generated inventions.

In February 2024, the U.S. Patent and Trademark Office (USPTO) published new guidance focused on the inventorship analysis for inventions assisted by AI. The guidance explicitly states that while AI-assisted inventions are not categorically unpatentable, the determination of inventorship must center on human contributions. This approach reflects the fundamental purpose of patents: to incentivize and reward human ingenuity.

The USPTO’s guidance emphasizes that a natural person must have made a “significant contribution” to the invention, even if AI systems were used during the inventive process. Although the guidance is not legally binding, it provides inventors, patent practitioners, and stakeholders with a framework to anticipate how the USPTO will evaluate patent applications involving AI and clarifies ongoing disclosure obligations before the Office.

This guidance was issued pursuant to the October 2023 Biden administration Executive Order on the Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence. Section 5.2(c)(i) of the Executive Order directed the USPTO to issue guidance addressing inventorship in the context of AI and to provide illustrative examples demonstrating how AI factors into the inventorship analysis.

Prior to this guidance, the Federal Circuit had acknowledged the open question of whether inventions created by humans with AI assistance are patent eligible. In Thaler v. Vidal, 43 F.4th 1207 (Fed. Cir. 2022), the court held that only natural persons can be inventors under 35 U.S.C. § 100(f), which defines “inventor” as an individual or individuals who invented or discovered the subject matter. The Federal Circuit reasoned that “individual” refers exclusively to human beings, consistent with Supreme Court precedent, and rejected petitions to name an AI system as an inventor.

The USPTO guidance clarifies that although AI systems cannot be named as inventors, the involvement of AI does not preclude patentability if a human inventor’s contribution meets the significant contribution standard.

To define “significant contribution,” the guidance references the Federal Circuit’s criteria from Pannu v. Iolab Corp., 155 F.3d 1344 (Fed. Cir. 1998). According to Pannu, each inventor must (1) contribute significantly to the conception or reduction to practice of the invention; (2) make a contribution that is not insignificant in quality relative to the entire invention; and (3) do more than merely explain well-known concepts or the current state of the art to the actual inventors.

The USPTO instructs that these factors apply to each claim in a patent application involving AI assistance. While an inventor need not contribute to every claim, each claim must have at least one human inventor who made a significant contribution. Given the increasing use of AI in invention creation, applicants are advised to exercise particular care in ensuring compliance with these standards.

The guidance sets forth five key principles to assist applicants in evaluating inventorship when AI is involved:

1. The use of AI does not negate a person’s status as an inventor. Simply employing AI tools does not automatically disqualify a human from inventorship.

2. Merely presenting a problem to an AI system is generally insufficient to establish inventorship. However, significant contribution may arise from the process of constructing prompts designed to elicit specific solutions from the AI, distinguishing this from the U.S. Copyright Office’s position that prompt entry alone is usually insufficient for authorship claims.

3. Merely reducing an invention to practice—such as implementing an AI-generated output—is not a significant contribution. A person cannot claim inventorship solely by applying AI output whose utility is apparent to skilled artisans.

4. A person who develops an essential building block, such as designing or training an AI system for a particular purpose, may be considered an inventor of inventions derived from that system.

5. Ownership or oversight of an AI system does not confer inventorship unless accompanied by a significant contribution to the invention’s conception.

The common thread across these principles is the requirement for active human participation in the inventive process, distinct from the autonomous operation of AI.

To illustrate these principles, the USPTO provided hypothetical examples. In one, a human used an AI-generated schematic to build a transaxle but substituted materials not specified by the AI. The USPTO concluded the human was not an inventor, as this amounted to mere reduction to practice without significant contribution. Conversely, in another example, human inventors made substantial contributions to the transaxle invention and then used AI to suggest manufacturing materials. Incorporating the AI’s suggestion into the claims did not negate their inventorship.