Navigating the Intersection of AI and Patent Law: Understanding the USPTO's Guidance on Generative AI Tools
"Navigating the Intersection of AI and Patent Law: Understanding the USPTO's Guidance on Generative AI Tools"
David Jackrel, PhD
Ask an AI chatbot like ChatGPT-4 about using generative AI to write a patent and it will proudly summarize the benefits of using AI, such as speed, efficiency, and consistency. It will also emphatically tell you that there are a myriad of potential risks, and that a legal professional should be consulted to maintain legal and technical accuracy, and overall quality. In direct opposition to the AI-taking-over-the-world scenario, today’s incarnation of AI seems to be humble and comes with its own warning label. The humility in this case reflects the reservations expressed by the majority of documented opinions, which were likely used to train the AI.
The USPTO officially weighed in on the matter, in the Guidance on Use of Artificial Intelligence-Based Tools in Practice Before the United States Patent and Trademark Office. The notice describes that existing rules are considered sufficient to “protect the USPTO's ecosystem” from the perils of such AI-based tools. The new guidance reminds inventors and patent practitioners (Patent Attorneys and Patent Agents) about which rules are pertinent, and about “the risks associated with the use of AI systems.”
The advent of impressive LLM-based tools (or “chatbots”) like Chat GPT-4 has caused an explosion of AI tools for professional use. IP is no exception, and many new tools are being developed and released for IP. The USPTO guidance confirms that AI is already being used by Examiners for prior art searches “using features like More Like This Document (MLTD) and Similarity Search in the Office's Patents End-to-End (PE2E) Search tool.” Likewise, the notice acknowledges that AI-based tools are available for practitioners, including for prior art research, patent application reviewing, drafting documents for submission to the USPTO, and gaining insight into examiner behavior.
A rule that governs the use of AI is the Duty of Candor and Good Faith, which includes a “duty to disclose to the Office all information known to that individual to be material to patentability.” This is relevant to AI, for example, where “those involved in patent proceedings have a duty to disclose all information—including on the use of AI tools by inventors, parties, and practitioners—that is material to patentability.” On the subject of using AI to generate an invention, the USPTO issued inventorship guidance and examples for AI-assisted inventions, which clarified that “AI-assisted inventions are not categorically unpatentable,” however, a human needs to contribute significantly to the invention for it to be patent eligible and the human is listed as the inventor, not the AI. However, the guidance also clarifies that “there is not presently a general duty to inform the USPTO that an AI tool was used in the drafting of the paper unless specifically requested by the USPTO.”
The USPTO Signature Requirement and Corresponding Certifications are also relevant to the use of AI. By signing a document or piece of correspondence with the USPTO, the signer is certifying that the statements made therein are believed to be true and a “duty of reasonable inquiry” is also imposed. The guidance specifically states that “simply relying on the accuracy of an AI tool is not a reasonable inquiry.” It goes on to explain that “the signature requirement and corresponding certifications ensure that documents drafted with the assistance of AI systems have been reviewed by a person and that person believes everything in the document is true and not submitted for an improper purpose.”
Even confidentiality and export regulations can become relevant when sending confidential information to an AI system that uses servers outside of the U.S. to process the information.
Generative AI is a powerful tool to produce written documents, but patents and other correspondence with the USPTO are highly critical and valuable assets that present unique challenges and risks. Patent practitioners are required to “reasonably consult with the client about the means by which the client's objectives are to be accomplished,” and the guidance from the USPTO is that practitioners should inform clients if they are using AI in the course of their work.
To sum it up, the notice offers the logical argument, “practitioners must competently represent their clients. That is, they must have the requisite legal, scientific, and technical knowledge to reasonably represent their client.” AI and generative AI are powerful tools with great potential to improve efficiency, but at the end of the day, the practitioner needs to be technically and legally qualified to represent the client, because, according to both the USPTO and the almighty data set that trained the AI, the AI is not.
© MLO, a professional corporation, 2024
**Disclaimer: This post is intended for informational purposes only and does not constitute legal advice. The information presented here may not apply to your specific circumstances. Consult with a qualified legal professional before making any decisions related to the topics discussed.**