AI IP Year in Review – AI Developments at the USPTO

Navigating Inventorship of AI-Assisted Inventions: USPTO’s Guidance and Implications

By Emily Tkac

The USPTO issued guidance on AI-assisted inventions on February 13, 2024. This guidance is part of the USPTO’s ongoing efforts to address the intersection of artificial intelligence (AI) and patent law. We focus on three main considerations regarding inventorship when AI systems are involved in the creation of inventions.

First, the USPTO clarifies that while AI-assisted inventions are patentable, AI systems themselves cannot be designated as inventors. Only natural persons can be inventors, which is in line with the Federal Circuit’s holding in Thaler v. Vidal, 43 F.4th 1207, 1213 (Fed. Cir. 2022).

Second, a person must make a “significant contribution” to the invention to qualify as an inventor of an AI-assisted invention. This contribution could involve designing or training the AI system, formulating prompts that lead to particular solutions, or enhancing the AI’s output through experimentation. However, merely recognizing a problem or appreciating the AI’s output does not qualify as inventorship. The Pannu¹ factors are the standard for determining whether a person has made a significant contribution to the invention. The Pannu factors assess the person’s role in conception, the quality of their contribution, and whether their input goes beyond explaining known concepts and the state of the art. Each inventor must satisfy each Pannu factor for each claim.

Third, the guidance reiterates that the duty of disclosure and reasonable inquiry still applies to AI-assisted inventions. Applicants must disclose all relevant information that might impact the patentability of the invention, including details about inventorship and contributions made by AI systems.

Overall, the USPTO’s guidance provides critical insights into how AI-assisted inventions will be treated under patent law, influencing both inventors and legal practitioners as AI technology continues to advance.

[1] Pannu v. Iolab Corp., 155 F.3d 1344, 1351 (Fed. Cir. 1998).

Update on USPTO Guidance for AI Subject Matter Eligibility

By: Todd Hopfinger, Lestin Kenton, Jr.

On July 16, 2024, the USPTO released updated guidance on patent subject matter eligibility for AI inventions. We focus on four main aspects of the updated guidance.

First, the updated guidance builds on previous guidance from 2019 and provides additional clarity on applying 35 U.S.C. § 101 to AI-related claims. It aims to ensure consistency in evaluating patent applications and patents involving AI.

Second, the updated guidance provides additional clarity around Step 2A of the USPTO’s eligibility analysis, focusing on whether a claim is directed to a judicial exception, such as abstract ideas, and if it integrates the exception into a practical application. On this front, the updated guidance incorporates stakeholder feedback and introduces three examples to demonstrate the application of the updated guidance to hypothetical AI-related claims.

Third, the updated guidance emphasizes that the framework for subject matter eligibility remains unchanged, and reassures stakeholders that AI inventions can be patented but that they should be carefully evaluated to avoid being dismissed as directed to a judicial exception under Step 2A of the USPTO’s eligibility analysis.

Finally, the updated guidance suggests that companies should be aware that while the USPTO recognizes the significance of AI technology and that AI inventions may be patented, the subject matter eligibility analysis framework for evaluating AI technology will likely continue to evolve both at the USPTO and in the courts.

USPTO Seeks Public Comment on the Impact of Artificial Intelligence on Patentability

By Roozbeh Gorgin, Ivy Estoesta

In early 2024, the USPTO issued a Request for Comments (RFC) seeking public input on the potential impact of AI on prior art, the knowledge of a person having ordinary skill in the art (PHOSITA), and determinations of patentability. See 89 FR 34217, p. 34217, available here (last visited June 10, 2024). The RFC is part of an ongoing initiative by the USPTO to explore the impact of AI technologies for patent applications, patent owners, patent practitioners, and the future of IP law.

The RFC sought input on 15 key questions, examining whether AI-generated disclosures should be treated as operable prior art, how AI impacts the skill level attributed to a PHOSITA, and whether current patent law accommodates AI’s complexities. The questions highlight concerns about AI’s ability to create vast disclosures without human oversight and AI’s potentially levelling effect on the “level of ordinary skill in the art,” potentially complicating determinations of novelty and nonobviousness.

Stakeholders’ comments, which were due by July 29, 2024, can now be viewed on the Regulations.gov website.

The USPTO has not yet issued guidance on how it will treat AI-generated prior art. However, that guidance (or lack thereof) has the potential to impact various stakeholders. For example, patent applicants and practitioners may face challenges in overcoming AI-generated prior art if their applications lack robust experimental data, resulting in longer prosecution times. Patent owners could see increased risks of invalidation due to the proliferation of AI-created references.

The USPTO’s inquiry underscores the evolving relationship between AI and patent law, signaling potential regulatory updates. Stakeholders are urged to engage actively, as the agency shapes a future that balances AI’s innovation potential with a resilient IP system.

Subject Matter Eligibility of AI Innovations—Updated Examples

By Todd Thurheimer

In July 2024, the USPTO issued guidance regarding the subject matter eligibility of patent claims involving AI. The guidance: (1) reaffirmed that the existing patent eligibility guidance framework applies to AI, (2) emphasized the potential for “practical applications” of AI, and (3) provided new Examples 47-49.

First, the guidance reaffirmed that the existing PEG framework will continue to be used to analyze claims across all technologies, including AI. The USPTO has published past examination guidance for subject matter eligibility to the MPEP (§ 2103-2106.07) and issued Examples 1-46¹ as resources for practitioners and examiners. These resources may remain highly relevant, useful tools for patent practitioners seeking protection of AI inventions.

Second, the guidance advised that Prong Two of Step 2A is a focal point for AI inventions, emphasizing the potential for practical applications of AI. A subject matter eligible claim should “cover a particular way to achieve a desired outcome, as opposed to merely claiming the idea of a solution or outcome.”

Third, the guidance was accompanied by new examples 47-49. Each example provides background information about the nature of the invention and the technology in place of a full specification, figures, etc. that would comprise an actual patent application. Each example includes a claim that is subject matter eligible and, as a point of contrast, a claim that is not.

The guidance is a part of ongoing efforts by the USPTO to clarify issues related to AI and provide guidance on subject matter eligibility. The intersection of these areas will continue to be a focal point of the USPTO in effectively promoting innovation, competition, and collaboration in burgeoning AI-based technologies and industries.

[1] The examples are available at www.uspto.gov/PatentEligibility. Many of the examples are based upon past Federal Circuit decisions given the USPTO’s ongoing efforts to monitor developments in the courts.

These articles appeared in the 2024 AI Intellectual Property Year in Review: Analysis & Trends report.