USPTO Guidance Applies Current Rules to AI
In April 2024, the United States Patent and Trademark Office issued guidance on use of artificial intelligence tools.
According to an article on the USPTO website announcing the guidance, the office hopes “to ensure AI is not misused – or left unchecked – in practice before our judicial boards.”
For those accustomed to practicing before the USPTO, the guidelines largely reflect common sense interpretations of existing USPTO rules. For example, they discuss the necessity of verifying the accuracy of AI-assisted documents submitted before the USPTO. The current AI guidance did not propose additional AI-specific rules.
The purpose of new USPTO guidelines
The guidance memo recognized the benefits of AI tools in IP prosecution and noted some of the ways that the USPTO is using AI, such as by USPTO patent examiners in prior art searches. “These tools have the potential to lower the barriers and costs of practicing before the Office as well as helping law practitioners offer services to their clients with improved quality and efficiency,” the memo stated.
On the other hand, the use of AI tools in the legal setting raises questions about accuracy, confidentiality, and legal ethics. Below, we discuss the existing rules and some of the specific warnings and examples the USPTO provided in the new memo, as well as analysis of the larger context for the new guidelines and other applicable rules.
Signature requirement and verification
USPTO rules currently require a signature by the person filing a document attesting to its truth. The new guidance does not change this requirement but points out that all statements must be true to the best of the person’s knowledge following a reasonable inquiry. While these are not new requirements, the use of AI tools amplifies the underlying concerns about knowledge and accuracy.
Verification: “[P]arties are responsible for the contents therein,” the guidance states. “Simply relying on the accuracy of an AI tool is not a reasonable inquiry.” The guidance warns against AI hallucinations. For example, an attorney should check the accuracy of citations and arguments.
The guidance notes, however, that practitioners are not currently required to inform the USPTO that an AI tool was used in drafting a document unless the USPTO specifically inquires. This differs from some courts and individual judges that have issued orders regarding if and how parties appearing before them are required to disclose AI use and/or certify the accuracy of AI-generated documents.
Signature: The guidelines also note that the signature of an AI tool “or other non-natural person” is not acceptable. The Trademark Manual of Examining Procedure contains additional information about who is authorized to sign particular documents. Any verification must be signed by 1) someone authorized to bind the trademark owner (such as a corporate officer), 2) an individual with firsthand knowledge and actual or implied authorization of the trademark owner, or 3) a qualified U.S. attorney with a power of attorney from the owner.
Evidence: The guidelines warn against submitting AI-generated evidence of trademark use “or any other evidence created by AI that does not actually exist in the marketplace.” This is consistent with other recent USPTO efforts to curb fabricated evidence of use and other documents.
The guidelines also state that “AI-generated material that misstates facts or law, includes irrelevant material, or includes unnecessarily cumulative material, could be construed as a paper presented for an improper purpose because it could ‘cause unnecessary delay or needless increase in the cost of any proceedings before the Office.’”
The Trademark Trial and Appeal Board Manual of Procedure similarly warns against irrelevant and/or cumulative material. While parties must be diligent in producing relevant evidence in inter partes Board matters, the Board looks with disfavor upon parties submitting voluminous records that force the Board to “spend needless time sifting through an inappropriately large record in search of germane proofs.” It is easy to imagine the record of a case becoming even larger when AI tools are used in discovery. Parties should be mindful of the proportionality of the volume of the evidence to the case.
Confidentiality
Under the USPTO rules as well as American Bar Association model rules as adopted by the USPTO, practitioners must take steps to maintain confidentiality and avoid inadvertent disclosure of confidential client information. Using new and developing AI tools elevates this concern. The USPTO recognizes that practitioners rely upon third-party tools, but notes that attorneys should be “especially vigilant” to ensure confidentiality. This could include breach of confidentiality for use in training data, national security concerns by use of AI tools based outside the U.S., and more.
Access to USPTO accounts
The guidance also notes that access to USPTO accounts, including trademark and patent filing systems, are exclusive to the registered user.
The guidelines warn against using AI tools to data mine information from USPTO databases, noting that using computer tools to generate unusually high numbers of database accesses violates USPTO terms of use and could even subject violators to state and federal criminal laws. Instead, the USPTO has made bulk data available through other outlets.
Duties to clients
While AI holds promise for the ability to increase access to legal services, there is also a danger of these tools being used incorrectly. Under the USPTO Rules of Professional Conduct, practitioners must have “the legal, scientific, and technical knowledge, skill, thoroughness and preparation reasonably necessary for the representation.” They also are required to consult with the client about the means used to accomplish the client’s objectives.
In AI context, the guidance states that attorneys must have the knowledge and skills to represent their clients. Practitioners cannot simply rely on AI. This is consistent with earlier ethical guidelines like the ABA Model Rules on competence to use other tools in electronic discovery.
AI and patent procedure
While this article focuses mainly on trademark practice, the USPTO guidance addresses patent-specific concerns as well. For example, the guidelines note that practitioners should be careful about questions of inventorship. Presently, U.S. law requires human inventorship for a patented invention, and attorneys should evaluate whether use of AI tools will affect this.
As generative AI tools continue to evolve, we will be watching for further guidance and developments from the USPTO.
For attorneys and brand owners interested in a secure, proprietary platform that has been helping companies and law firms of all sizes for more than two decades and continues to evolve with new technologies and customer needs, learn more about WebTMS.
About the Author: Suzi Morales is a writer and editor specialising in IP. She previously practised IP law, including trade mark prosecution and all types of IP litigation, for 15 years. Learn more about Suzi’s work at www.suzimorales.com/intellectual-property-law-marketing