Generative AI in USPTO Practice: Key Considerations Under the USPTO’s New Guidance

Generative Artificial Intelligence (AI) tools such as OpenAI’s ChatGPT, Anthropic’s Claude, Midjourney and others, have made waves across various industries and the legal profession is no exception. Specialized publications and recent news are replete of examples of professionals in all fields increasingly turning to these tools to streamline their work.

However, the use of generative AI in legal practice also raises special concerns about the potential for errors, bias, and ethical violations. In recent high-profile cases, the use of ChatGPT by attorneys came under scrutiny when their court filing were found to contain false statements and references to non-existent legal authorities. In one case, two lawyers were sanctioned for submitting non-existent AI-generated judicial opinions with fake quotes and citations, without properly verifying the accuracy of such citations. See Mata v. Avianca., No. 22-CV-1461 (PKC), 2023 WL 4114965 (S.D.N.Y. June 22, 2023). This incident highlights the need for lawyers to exercise caution and maintain human oversight when using generative AI tools.

Building upon the growing awareness of the pervasive use of generative AI, and its gradual adoption in the legal profession, the United States Patent and Trademark Office (USPTO) has taken steps to address the use of AI tools in practice before the agency. In February 2024, USPTO Director Katherine K. Vidal issued a memorandum to the Patent Trial and Appeal Board (PTAB) and Trademark Trial and Appeal Board (TTAB), emphasizing the applicability of existing regulations to party and practitioner misconduct related to the use of AI.

Director Vidal’s memorandum highlighted the potential benefits of AI in increasing access to information and lowering costs, while also acknowledging the risks associated with AI-generated content, such as inaccuracies and “hallucinations.” The memo served primarily as a reminder that existing USPTO rules, including the Rules of Professional Conduct, apply to AI-assisted submissions and that practitioners must exercise due diligence in reviewing and verifying the accuracy of such content.

Following the February memorandum, the USPTO published a more comprehensive guidance notice in the Federal Register on April 11, 2024, titled “Guidance on Use of Artificial Intelligence-Based Tools in Practice Before the United States Patent and Trademark Office.” This notice expands upon the principles outlined in Director Vidal’s memo and provides a detailed framework for the responsible use of AI tools in patent and trademark matters.

Some of the key points from this notice include:

• The USPTO recognizes the possibility that AI will be used to prepare and prosecute patent and trademark applications, as well as other filings before the Office including filings submitted to the Patent Trial and Appeal Board (PTAB) and Trademark Trial and Appeal Board (TTAB)
• There is no prohibition against using AI tools in drafting documents for submission to the USPTO, nor a general obligation to disclose the use of such tools.
• Existing USPTO rules and policies apply broadly to situations where a party or practitioner uses AI as a tool, regardless of AI assistance in preparing submissions. Practitioners must ensure statements are true to their own knowledge, that arguments are legally warranted, and facts have evidentiary support.
• Parties must review and verify the contents of AI-assisted documents. Simply relying on the accuracy of an AI tool is not considered a reasonable inquiry.
• Nearly all forms of correspondence with the USPTO must be signed. This includes documents that were drafted entirely by AI tools or drafted with the assistance of AI tools. Signatures on USPTO correspondence must be personally entered by the named signatory and cannot be delegated to an AI tool. The signatory is certifying they reviewed the document.
• Authorized persons must ensure AI tools do not exceed the person’s authorized access. It is therefore not acceptable for the correspondence to have the signature of an AI tool or other non-natural person.
• AI tools may be used to assist or automate the mechanical aspects of filing documents with the USPTO. For example, these tools can be employed to autocomplete USPTO forms and upload documents and other information to USPTO servers. However, AI tools may not be associated to a USPTO.gov account for filing and/or accessing documents electronically via the USPTO’s electronic filing system.
• Fraud or misconduct through the use of AI tools in USPTO proceedings or systems is not tolerated. Existing duties of candor and good faith apply to an individual’s actions taken with AI.
• Use of AI systems to perform prior art searches, application drafting, etc. may result in the inadvertent disclosure of client-sensitive or confidential information to third parties through the owners of these systems, causing harm to the client. Practitioners must take steps to maintain confidentiality since using AI tools could inadvertently disclose client-sensitive information to third parties.

The USPTO’s guidance on the use of AI generative tools in practice before the agency is not the end of the conversation – far from it. It effectively invites the parties and their counsel to engage more deeply with the complex issues surrounding generative AI. At the same time, the agency demonstrates its openness to the adoption of this transformative technology, while prioritizing the integrity of the intellectual property system and the ethical obligations of USPTO practitioners.

Gibbons will continue to monitor and report developments in the areas of AI and Intellectual Property law.