The United States Patent and Trademark Office (“USPTO”) issued guidance on February 13, 2024 (the “Guidance”) regarding the patentability of inventions created or developed with the assistance of artificial intelligence (“AI”), a novel issue on which the USPTO has been seeking input from various public and private stakeholders over the past few years.  President Biden mandated the issuance of such Guidance in his executive order on AI (see our prior alert here)[1] in October 2023.  The Guidance aims to clarify how patent applications involving AI-assisted inventions will be examined by patent examiners, and reaffirms the existing jurisprudence maintaining that only natural persons, not AI tools, can be listed as inventors.  However, the Guidance clarifies that AI-assisted inventions are not automatically ineligible for patent protection so long as one or more natural persons “significantly contributed” to the invention.  Overall, the Guidance underscores the need for a balanced approach to inventorship that acknowledges both technological advancements and human innovation.  The USPTO is seeking public feedback on the Guidance, which is due by May 13, 2024.

The Guidance builds on cases challenging the patentability of AI-assisted inventions; in particular, Thaler v. Vidal,[2] in which the Federal Circuit reaffirmed the traditional understanding of inventorship, holding that AI tools cannot themselves be listed as inventors.  However, the Thaler court did not elucidate the contours of what degree or type of human contribution is required to constitute inventorship where use of AI tools is involved.  The Guidance builds on this jurisprudential framework and provides further instruction to inventors, examiners and other stakeholders on how to determine whether a natural person’s contribution to an invention rises to the threshold of inventorship.

Outlined in the Guidance are a number of non-exhaustive principles for assessing human contributions to AI-assisted inventions.  Although the Guidance points out that there is no bright line test for determining a person’s contribution, it repeatedly points to the Pannu[3] factors, historically relied on by courts to evaluate joint ownership, as relevant considerations for assessing human contributions to AI-assisted inventions and evaluating when a natural person has made a “significant contribution” sufficient to be considered an inventor.  For instance, the USPTO suggests, in light of the Pannu factors, that a person who merely recognizes an output of an AI system, particularly when the properties and utility of the output are apparent to those of ordinary skill, may not qualify as an inventor, but “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.”  The Guidance addresses the scenario where an AI tool answers a question posed by a human.  In this case, simply posing a problem to an AI system is unlikely to constitute a significant contribution by the human.  However, the Guidance proposes that where a natural person develops an “essential building block” from which an invention stems, such person may be considered to have provided a significant contribution to the conception of the invention, .  The Guidance explains:  “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.”  The Guidance also notes that simply maintaining “intellectual domination” over an AI system, or overseeing an AI system used in the creation of an invention, does not make that person an inventor. 

Highlights of the Guidance

To recap, key points highlighted in the Guidance include:

  1. Clarity on “Significant Contribution”:  The USPTO emphasizes that AI is merely a tool used by humans and cannot not constitute an inventor.  Therefore, patent applications must clearly attribute inventive concepts to human inventors to be eligible for patent protection.  However, a natural person’s use of AI does not negate the person’s contributions as an inventor, and such person may be listed as an inventor if the natural person “contributes significantly” to the AI-assisted invention.  Factors such as conception[4] and the quality of human involvement are crucial aspects in determining inventorship.  Every claim of a patent must have a human inventor. 
  1. Disclosure Requirements:  The Guidance reiterates that patent applicants owe the USPTO a duty of candor and good faith in dealing with the USPTO, which includes a duty to disclose all known information that is material to patentability, or, conversely, a duty to disclose “information that raises a prima facie case of unpatentability due to improper inventorship,” such as evidence that a named inventor did not significantly contribute to the invention, but rather, such contribution was made by an AI tool.
  1. Case-by-Case Evaluation:  The Guidance stresses that determinations of inventorship for AI-assisted inventions will be made on a case-by-case basis in view of the specific facts of the application, including extrinsic evidence on the record.  Examiners and other USPTO employees are able to request additional information as may be reasonably necessary to properly examine an application, and applicants may see an increase in office actions as a result of the Guidance.  This approach allows for a nuanced assessment of each patent application’s unique circumstances, ensuring fairness and accuracy in the evaluation process.
  1. Examiner Training:  The USPTO plans to provide additional training to patent examiners to ensure consistent and accurate evaluation of patent applications involving AI.

Comparison with USCO Guidance on Copyrightability of AI-Generated Works

Interestingly, the position taken by the USPTO in the Guidance diverges somewhat from that taken by the U.S. Copyright Office (“USCO”) with respect to the copyrightability of works created or generated with the assistance of AI.  In March 2023, the USCO issued its own statement of policy, taking the view that AI-generated material, even if containing expressive elements or influenced by human prompting, is not protected by copyright for lack of human authorship (though, the USCO points out, if a human were to, for example, arrange the AI-generated material in a sufficiently creative way, the resulting human-authored elements may qualify for protection as an original work of authorship).  Subsequently, the U.S. District Court for the District of Columbia confirmed in Thaler v. Perlmutter that works developed exclusively and autonomously by an AI tool are not copyrightable for lack of human authorship, but left open the question as to what level of human contribution or creativity may suffice to entitle an AI-assisted work to copyright protection.[5]  Additionally, whereas the USPTO, as mentioned above, will not require patent applicants to disclose whether AI tools were used or contributed to the invention so long as the application attributes proper human inventorship to each claim of the applied-for patent, the USCO clarified in its 2023 guidance that copyright registrants must (1) include a description of any AI-generated content included in the work and (2) exclude any AI-generated material that is more than de minimis from the copyright application.  Its guidance also instructed on how to draft a copyright registration to effectively disclose the use of generative AI tools, and how to amend a previously submitted pending application to correct any disclosure deficiencies.  Notwithstanding these distinctions, the USPTO and USCO are aligned in their respective views on how to evaluate inventorship/authorship when AI is involved, both reaffirming that patent and copyright protection only extends to the ingenuity and creativity of natural persons.  Further, both agencies put significant emphasis on the inaptness of a bright line test and the need to consider these questions on a case-by-case basis.

Overall, the Guidance provides valuable clarity and direction for promoting innovation and navigating the complexities of seeking patent protection for AI-assisted inventions, while upholding the integrity of the patent system and U.S. intellectual property regime at large. 

[1] Exec. Order. No. 14110, 88 Fed. Reg. 75191 (Nov. 1, 2023).

[2] 558 F.Supp.3d 238 (E.D. Va. 2021).

[3] Pannu v. Iolab Corp., 155 F.3d 1344 (Fed. Cir. 1998).  The Pannu factors include whether the inventor
(1) contributed in some significant manner to the conception or reduction to practice of the invention; (2) made 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) did more than merely explain to the real inventors well-known concepts and/or the current state of the art.

[4] Note, the USPTO points out that, while the Pannu factors also refer to reduction to practice, reduction to practice alone is insufficient to demonstrate inventorship (citing Section 2109 of the Manual of Patent Examining Procedure regarding inventorship).  The USPTO highlights that Pannu only acknowledges that the doctrine of simultaneous conception and reduction to practice is relevant in “unpredictable arts” where the conception may not occur until reduction to practice occurs.  See, e.g., Amgen, Inv. V. Chugai Pharm. Co., 927 F.2d 1200, 1206 (Fed. Cir. 1991).

[5] See our blog post on the Thaler v. Perlmutter decision regarding the copyrightability of AI-generated works here:  Paradise Lost: Human Involvement Still Required for “Creativity Machine’s” Work.