Insights into employing generative AI as directed by the USPTO

The USPTO has just released two new announcements concerning generative AI – one focused on submission preparation and the other dealing with ownership issues. in this episode, we'll delve into both topics in detail.

Read: Patent Practitioners Take Note: (Even) if You Use AI, Read Before You File

Read: USPTO AI GUIDANCE: Human vs. Machine – Humans Win … For Now

About the authors:
Samuel A. Savanich
Alexander P. Wharton
Dr. Christian E. Mammen
Jeffrey S. Whittle
Stephanie Smiertka Riley
Michael J. Mallie

Transcript

Welcome to Womble Perspectives, where we explore a wide range of topics, from the latest legal updates to industry trends to the business of law. Our team of lawyers, professionals and occasional outside guests will take you through the most pressing issues facing businesses today and provide practical and actionable advice to help you navigate the ever changing legal landscape.

 With a focus on innovation, collaboration and client service. We are committed to delivering exceptional value to our clients and to the communities we serve. And now our latest episode.

On today’s episode, we’re covering two recent developments pertaining to the USPTO and generative AI.  On February 6th and February 13th, the USPTO issued guidance on using generative AI in preparing submissions and guidance on ownership, respectively. 

 First, let’s take a look at using generative AI in preparing submissions:

 On February 6, 2024, the USPTO weighed in with guidance for practitioners using—or considering using—AI in preparing submissions to the USPTO.  In essence, Director Kathi Vidal has reminded practitioners that they must sign their submissions, that the signature constitutes a certification of its accuracy, and that the signers are therefore responsible for the accuracy of those submissions, including anything AI-generated. This guidance precedes the expected Federal Register guidance likely to be published at the end of February. 

 This clarification explained the USPTO’s position on the use of AI (including generative AI content) in legal proceedings and directed the Patent Trial and Appeal Board and Trademark Trial and Appeal Board to hold parties responsible for the misuse of AI in legal proceedings.

 Director Vidal offered her comments in light of the discussion of AI by Chief Justice John Roberts in the Year-End Report on the Federal judiciary released on December 31, 2023. The inevitability of AI in USPTO practice also is bolstered by President Biden’s October 30th Executive Order. While the forthcoming Federal Register guidance will address “inventorship and the use of AI… in the inventive process, per the Order, the guidance this week was directed to the Patent Trial and Appeal Board and Trademark Trial and Appeal Board, expecting staff to “apply their existing skills and relevant existing rules to the challenges the Chief Justice identified.” 

 The Chief Justice recognized the potential use for AI, as well as the potential pitfalls, referencing recent court filings generated via AI including “citations to non-existent cases” (referred to as “AI hallucinations”). Director Vidal noted that, under the current USPTO Rules of Professional Conduct, any submission to the USPTO under signature must be reviewed for accuracy by the person presenting that submission. For example, such review includes recognizing errors or omissions, as well as verifying factual and legal accuracy. As reiterated by Director Vidal, such clarifications apply not only to AI generated documents, but to any document submitted to the USPTO.  

 She wrote that “practitioners are also prohibited from asserting or controverting an issue in a proceeding unless there is a basis in law or fact for doing so,” another potential problem with the submission of unchecked AI-generated documents. “Simply assuming the accuracy of an AI tool,” writes Director Vidal, “is not a reasonable inquiry.”

 Director Vidal further noted the potential consequences of including such inaccuracies or errors. For example, the submission could be struck or a practitioner could be precluded from submitting a paper. Additional and more severe consequences exist, such as terminating the proceedings in the Office and, potentially, criminal liability and disciplinary action. 

So what does this mean and how could this affect you? These clarifications should not significantly impact current practice. It is good practice to always review, for any situation, rather than just for legal documents or submissions, any AI-generated document. Further, any document, whether generated by AI or not, prior to submission to the USPTO, should be thoroughly reviewed with adequate time and care, despite the time pressures practitioners typically face. Although AI, particularly generative AI, may be useful for quickly solving problems and/or generating drafts of documents, the contents of those documents should never be taken as the truth until reviewed for accuracy. Many examples have now been reported regarding “AI hallucinations” or, in other words, of AI fabricating facts, quotes, and cases.

 The takeaway? Like all matters related to generative AI, it's tread carefully and always inspect the content produced by AI.

 Next up: who wins in the contest between human  and machine when it comes to AI-assisted inventions:

 On February 13, 2024, the USPTO published a Federal Register notice on Inventorship Guidance for Artificial Intelligence-assisted Inventions in response to President Biden’s October 2023 Executive Order on the Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence. The USPTO has been grappling with the interplay of artificial intelligence and the patent system since August 2019, when it first issued a request for public commentary on the patentability of AI-assisted inventions. The guidance largely tracks the holding of a 2022 Federal Circuit decision, Thaler v. Vidal, which affirmed the district court and PTO decisions that individuals, as defined in the Patent Act, plainly mean humans and affirmed the denial of the USPTO application listing only an AI software system as an inventor. Despite this decision, the reality is that AI is being increasingly used during primary research. Because there is no legislation expressly addressing the role of AI in development of patentable inventions, practitioners look to agency guidance, such as this, for some answers. In the age-old battle of Human vs. Machine, this Guidance unequivocally shows that inventorship requires a human.  

 In this Guidance, the USPTO states that the use of AI does not preclude patentability, so long as at least one natural person significantly contributed to the claimed invention and can be named as the inventor. It is clear that this standard will entail some line drawing in the future, but the USPTO has directed courts to consider the Pannu factors, which is a three-part test for naming inventors that requires the inventors to have: “(1) contributed in some significant manner to the conception 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) [done] more than merely explain to the real inventors well-known concepts and/or the current state of the art.” Example 1 indicates that simply providing a query or prompt to an AI algorithm and utilizing the output as-is, does not equate to human inventorship. The person, to become an inventor, must contribute to the claims in some way, other than merely entering a query or prompt.

 Although the Guidance is helpful to know where the USPTO stands on the issue (with humans), there is now a new layer of complexity. Determining inventorship will still involve determining what a named (human) inventor contributed to a particular claim set, regardless of whether AI was utilized or not. However, any use of AI serves only to complicate that process. When inventive processes are interwoven with AI processes, applicants and their legal counsel will need to ensure that the inventor substantially contributed to the invention. Tracking inventorship is always recommended but will be particularly important moving forward for AI-assisted inventions. Despite speculation of a possible disclosure requirement, the USPTO confirmed it is not requiring affirmative disclosure that AI contributed to an invention. This differs from the U.S. Copyright Office’s policy which does require disclosure. However, applicants are still required to submit an oath or declaration of proper inventorship. In addition to the above, the Guidance provides five guiding principles to practitioners to further assist with the AI v. human inventorship analysis. Overall, this Guidance helps fill the void relating to AI and inventorship for now. However, as AI continues to seep into our daily lives, Congressional action may not be far off.

Thank you for listening to Womble Perspectives. If you want to learn more about the topics discussed in this episode, please visit The Show Notes, where you can find links to related resources mentioned today. The Show Notes also have more information about our attorneys who provided today's insights, including ways to reach out to them.

Don't forget to subscribe via your podcast player of choice so that you never miss an episode. Thank you again for listening.