The U.S. Patent and Trademark Office (USPTO) has issued guidance regarding patentability and inventorship for inventions made with the assistance of artificial intelligence (AI), clarifying the Office’s position that such inventions are patent-eligible so long as at least one human inventor made a significant contribution to conception of the invention.
In addition to announcing its view that AI-assisted inventions may be patent-eligible, the USPTO provided several guiding principles that it believes should inform determinations as to whether and when a human should be considered an inventor when using AI assistance. Finally, the USPTO provided two detailed examples to guide examiners, practitioners, and inventors in determining inventorship for AI-assisted inventions.
Background
The Federal Circuit’s 2022 decision in Thaler v. Vidal, 43 F.4th 1207 (Fed. Cir. 2022) established that only a human inventor—not an AI system—can be validly named as an inventor on a U.S. patent application. Thus, under Thaler, inventions that are invented entirely by AI systems are not eligible for U.S. patent protection. However, the Thaler court explicitly acknowledged that its decision did not confront the more nuanced question of “whether inventions made with the assistance of AI are eligible for patent protection.” Thaler v. Vidal, 43 F.4th at 1213.
Following the Thaler decision, the Biden administration’s AI Executive Order directed the USPTO director to weigh in on the very issue flagged as unresolved by the Thaler court. The AI Executive Order of Nov. 1, 2023, charged the USPTO director to “publish guidance to USPTO patent examiners and applicants addressing inventorship and the use of AI, including generative AI, in the inventive process.” Executive Order 14110, Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence, 88 Fed. Reg. 75191-75226 (Nov. 1, 2023).
The USPTO has now issued that guidance (Inventorship Guidance for AI-Assisted Inventions, 89 Fed. Reg. 10043-10051 (Feb. 13, 2024)) and is seeking public comment on it for a 90-day period.
USPTO Position: Significant Human Contribution to Conception Is Required
The standard endorsed by the USPTO in its guidance for AI-assisted inventions is the following: “While AI systems . . . cannot be listed as inventors on patent applications or patents, the use of an AI system by a natural person(s) does not preclude a natural person(s) from qualifying as an inventor (or joint inventors) if the natural person(s) significantly contributed to the claimed invention.” 89 Fed. Reg. at 10046.
Critically, significant contribution to an invention requires significant contribution to the conception of the invention. Traditionally, conception is defined as the formation of an invention in the mind of an inventor. Conception is distinct from reduction to practice, which involves the processes of actual construction, successful testing, validation, or demonstrating an invention’s workability. While reduction to practice may occur simultaneously with conception in some situations, significant contribution to reduction to practice alone is not enough for a human contributor to be an inventor if the human contributor did not contribute in any significant manner to conception.
Pannu Factors
To determine whether a human inventor has made a significant contribution to the conception of an invention, the USPTO looks to the Federal Circuit’s Pannu factors. (See Pannu v. Iolab Corp., 155 F.3d 1344, 1351 (Fed. Cir. 1998).) While the Pannu factors were promulgated by the Federal Circuit with respect to determining inventorship among multiple human contributors, the USPTO guidance says that these factors should apply similarly in determining inventorship among cooperative human and AI contribution.
The Pannu factors provide the following requirements of an inventor: “. . . (1) contribute in some significant manner to the conception or reduction to practice of the invention; (2) make a contribution to the claimed invention that is not insignificant in quality, when that contribution is measured against the dimension of the full invention; and (3) do more than merely explain to the real inventors well‑known concepts and/or the current state of the art.” Id. at 1351 (citing to Fina Oil Chem. Co. v. Ewen, 123 F.3d 1466, 1473, 43 USPQ2d 1935, 1941 (Fed. Cir. 1997); and Ethicon, Inc. v. United States Surgical Corp., 135 F.3d 1456, 1460, 45 USPQ2d 1545, 1548 (Fed. Cir. 1998)) (emphasis added).
Principles for Applying Pannu Factors
In addition to invocation of the Pannu factors, the USPTO provided several “principles” that it says should inform application of the Pannu factors in situations where AI is leveraged to assist in creation of an invention. These principles articulated in the USPTO guidance include the following:
- A natural person’s use of an AI system in creating an AI-assisted invention does not negate the person’s contributions as an inventor. The natural person can be listed as the inventor or joint inventor if the natural person contributes significantly to the AI-assisted invention.
- Merely recognizing a problem or having a general goal or research plan to pursue does not rise to the level of conception. A natural person who only presents a problem to an AI system may not be a proper inventor or joint inventor of an invention identified from the output of the AI system. However, 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.
- Reducing an invention to practice alone is not a significant contribution that rises to the level of inventorship. Therefore, a natural person who merely recognizes and appreciates the output of an AI system as an invention, particularly when the properties and utility of the output are apparent to those of ordinary skill, is not necessarily an inventor. However, a person who takes the output of an AI system and makes a significant contribution to the output to create an invention may be a proper inventor. Alternatively, in certain situations, a person who conducts a successful experiment using the AI system’s output could demonstrate that the person provided a significant contribution to the invention even if that person is unable to establish conception until the invention has been reduced to practice.
- A natural person who develops an essential building block from which the claimed invention is derived may be considered to have provided a significant contribution to the conception of the claimed invention even though the person was not present for or a participant in each activity that led to the conception of the claimed invention. In some situations, the natural person(s) 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.
- Maintaining “intellectual domination” over an AI system does not, on its own, make a person an inventor of any inventions created through the use of the AI system. Therefore, a person simply owning or overseeing an AI system that is used in the creation of an invention, without providing a significant contribution to the conception of the invention, does not make that person an inventor.
89 Fed. Reg. at 10048-10049 (emphasis added).
Human Inventorship Is Determined Claim by Claim
Consistent with the principle that inventorship among multiple human inventors is determined on a claim-by-claim basis, the USPTO guidance sets forth that human inventorship for AI-assisted inventions is also determined on a claim-by-claim basis.
The most interesting implication of this claim-by-claim principle with respect to AI-assisted inventions is that each claim must have a valid human inventor. As set out in some detail in the examples provided by the USPTO with its guidance, the implications for dependent claims are somewhat interesting. For example, a dependent claim that recites a minor contribution made entirely by AI may still be included in an application, so long as the underlying independent claim has a valid human inventor who made a significant contribution to the conception of the overall inventive concept.
Other Guidance Notes
Other points in the USPTO guidance that may be of interest to inventors and practitioners include the following:
- Design patents and plant patents – The USPTO’s guidance applies equally to design and plant patents as it does to utility patents.
- Duty of disclosure – The duty of disclosure for information material to patentability, under which anyone involved in preparation or prosecution of a patent application must submit material information to the Office in an Information Disclosure Statement, applies to information that is material to determining inventorship.
- Duty of reasonable inquiry – According to the USPTO’s guidance, practitioners may be required to inquire about whether and how AI is being used in the inventive process, in order to fulfill their duty of reasonable inquiry to obtain information that may be material to patentability (which may thus be required to be disclosed to the Office).
- Priority claims – While the U.S. courts have foreclosed the possibility that a patent application can validly name an AI system as a sole inventor or joint inventor, this is not the case in all countries. Thus, the USPTO guidance addresses the question of whether and how priority can be claimed in U.S. applications to foreign applications listing an AI inventor. According to the guidance, a U.S. application cannot claim priority to a foreign application that names an AI system as a sole inventor, but a U.S. application may claim priority to a foreign application that names humans and AI systems as joint inventors. The U.S. application must itself name only human inventors, and the claims of the U.S. application should be reviewed to ensure that the listed human inventor contributed in a significant manner to each of them.
USPTO AI-Assisted Invention Examples
In addition to the guidance memorandum itself, the USPTO also provided two detailed examples, outlining how it believes its AI-assisted invention principles should be applied. The first example explores situations in which a publicly available generative AI system is used to create a transaxle design for a remote-control car. The second example explores situations in which a ready-to-use deep neural network (DNN) prediction model is used, sometimes in combination with bespoke generative AI models, in discovery and development of therapeutic compounds.
A brief summary of the USPTO examples is provided below.
Example 1 – Transaxle for a remote-control car
In this example, toy car engineers prompt a publicly available generative AI system to create a transaxle design: “Create an original design for a transaxle for a model car, including a schematic and description of the transaxle.”
- Scenario 1 of Example 1:
- An independent claim recites the AI-generated transaxle design.
The USPTO guidance explains that listed human inventors are not proper because they did not make a significant contribution to the conception of the AI-generated invention.
- Scenario 2 of Example 1:
- A dependent claim, dependent from the independent claim reciting the AI-generated transaxle design, recites the use of steel as a material, which was selected by a human inventor.
The USPTO guidance explains that listed human inventors are not proper because selection of a common material is not a significant contribution to the overall claim scope.
- Scenario 3 of Example 1:
- Human engineers modify the AI-generated axle design, and an independent claim recites the modified axle design.
The USPTO guidance explains that a listed human inventor is proper because human engineers made significant contributions to conception by making significant alterations to the original AI-generated design.
- Scenario 4 of Example 1:
- A dependent claim, dependent from the independent claim reciting the modified, human-generated transaxle design, recites the use of aluminum as a material, which was conceived in part by the human engineer in reliance on an AI-generated suggestion.
The USPTO guidance explains that a listed human inventor is proper because human engineers made significant contributions to the conception of the underlying independent claim, which are not negated by the use of AI assistance in the conception of the dependent claim.
- Scenario 5 of Example 1:
- The creator and owner of the publicly available general-purpose generative AI system used by the remote-control car engineers does not otherwise participate in the inventive process, but is listed as an inventor on an independent claim.
The USPTO guidance explains that the AI engineer is not proper as an inventor in this case because creation of a general-purpose AI system (without designing, building, or training the AI system in view of a “specific problem” to elicit a “particular solution” in a manner that significantly contributes to the invention of the solution created with the AI system) does not alone constitute a significant contribution to solutions created using the AI system.
Example 2 – Developing a therapeutic compound for treating cancer
In this example, drug researchers use a ready-to-use DNN prediction model to select candidate compounds for cancer treatment. The researchers provide drug-target pairs as input in Simplified Molecular Input Line Entry System (SMILES) string format, and the prediction model outputs candidate compounds with the highest predicted binding affinity.
- Scenario 1 of Example 2:
- Wet-lab researchers manually identify and synthesize structural modifications to the AI‑identified candidate compounds.
- Independent claim 1 is directed to a method for identifying a compound comprising using the DNN, synthesizing an intermediate, and synthesizing a compound by introducing structural modifications.
- The USPTO guidance explains that the wet-lab researchers are proper inventors because of their significant contributions to the intermediate synthesis and creation of structural modifications, whereas the AI researcher who did imputation using the DNN is not proper as an inventor (as the use of the DNN was an insignificant contribution in quality when measured against the scope of the entire claim).
Independent claim 2 recites the structurally modified compound itself.
- The USPTO guidance explains that, similar to claim 1, the wet-lab researchers are proper inventors, whereas the AI researcher who did imputation using the DNN is not proper as an inventor.
- Scenario 2 of Example 2:
- AI researchers build, train, and fine-tune a new generative neural net system that accepts SMILES inputs and generates an optimized drug compound (not previously synthesized). The AI researchers apply the generative AI system downstream from the DNN prediction model to generate structurally modified versions of the candidate compounds identified by the DNN. Wet-lab researchers then synthesize the structurally modified compounds. An independent claim is directed to one such structurally modified compound.
The USPTO guidance explains that both the wet-lab researchers who synthesized the compounds and the AI researchers who built, trained, and fine-tuned the generative AI model in view of a specific problem to elicit a particular solution are proper inventors.
While the examples outline certain situations in which human beings have not made any significant contribution to the conception of a claimed invention, and therefore cannot be listed as valid inventors, the examples seem overall (particularly with respect to the second example relating to AI-assisted drug discovery) to seek to provide pathways by which inventors, practitioners, and examiners can establish valid human inventorship. The examples, however, still leave significant uncertainty as to what amount of human contribution is required to establish valid human inventorship. Notably, Example 2 is silent on any scenarios in which structural modifications are not made to a candidate compound identified or generated by a ready-to-use AI system, which may signal a scenario in which proper human inventorship of a claim to such a compound could be challenged.