Following AI guidance published earlier this year (USPTO Issues New Guidance on Inventorship when Assisted by AI), the US Patent & Trademark Office (USPTO) continues to update and expand its guidance concerning the use of AI tools by parties and practitioners. While previous guidance focused in on the impact of AI on development and inventorship (and the barriers that such use may present towards seeking proper patent protection), this most recent guidance turns to the use of AI technology to prepare and prosecute patent applications. While the USPTO understands the benefit and time-saving potential offered by AI, the new guidance reminds patent practitioners of existing rules and policies that may govern the use of AI when prosecuting patents and trademarks.
The new guidance references many existing ethical and professional responsibility requirements—including the duty of candor and faith, signature requirements, and confidentiality requirements—that may be implicated by the use of new AI technologies. The USPTO confirms that the material use of AI in the development of any inventions should be disclosed, but similar to past guidance, the exact instances where the use of AI may need to be disclosed are not explicitly defined. Practitioners will have to use their judgment, in combination with the guidelines set out by the USPTO, when advising clients on the use of AI as part of the inventive process.
Additionally, the new USPTO guidance addresses the use of AI by patent and trademark professionals in their work before the USPTO. For example, in recent years various AI tools have been marketed to draft patent specifications, generate responses to USPTO office actions, write and respond to briefs, and draft patent claims. The guidance confirms that there is no general prohibition against using these computer tools in drafting documents for submission to the USPTO nor is there an obligation for a practitioner to disclose such use to the USPTO. There is, however, a risk that haphazard use of AI tools when drafting correspondence to the USPTO may be drawing from unreliable sources of information and generate inaccurate or irrelevant outputs. The submission of such unreviewed materials may constitute a violation of the USPTO’s signature requirements, which require the applicant or their representative to sign correspondence to the USPTO certifying the truth of all statements made in the correspondence. It is therefore important for attorneys to review, verify, and correct (if necessary) the contents of correspondence and representations to the USPTO. Submissions to the USPTO should also not be automated, based on the above-mentioned signature requirements. Similarly, the USPTO points out that claims based, in any degree, on AI contributions should be modified to a presentable, patentable state. The failure to review correspondence, submissions and filings to the USPTO may also constitute a breach of a practitioner’s duty to provide competent and diligent representation on behalf of their clients.
The USPTO also offers more general commentary on the risks of using inadequately secured AI systems for matters before the USPTO. These risks stem from confidentiality and national security concerns regarding the use of open AI systems that may result in the inadvertent disclosure of sensitive materials. To the extent possible, practitioners should work with their clients to ensure that the use of AI is not leading to any improper disclosures of confidential information.
This is a fast-emerging area of law, with several state bar associations—including those in California, Florida, New Jersey, and New York—issuing guidelines in recent months setting forth similar guidance on the ethical use of AI. A number of courts and judges have also put standing orders in place to address the risks of using generative AI tools to prepare court filings. Patent practitioners should continue to monitor guidance from the USPTO closely and should stay abreast of their obligations with respect to the responsible use of AI.
Full text of the recent guidance can be located here: USPTO issues guidance concerning the use of AI tools by parties and practitioners | USPTO
