In the past year, I have seen an increase in questions related to artificial intelligence. Specifically, patent and trademark lawyers have asked whether it is permissible for those lawyers engaged in practice before the USPTO to use Generative AI. While I have and continue to present on this topic multiple times, including (here), (here), and (here), it is helpful to have a quick primer on the use of AI by IP lawyers.
First, it is helpful to know that Generative AI can be helpful in patent and trademark practice, including to assist lawyers in:
- Analysis of Office Actions
- Preparing Draft Patent Applications
- Invention Harvesting
- Legal Research
- General Analysis of Success on the Merits
- Review of Specimens
- Predictive analysis regarding refusals
- Analysis of subject matter Conflicts of Interest
However, with recent disciplinary cases, sanctions orders, and guidance from courts and the USPTO regarding AI, it is helpful to think through the implications under the USPTO Rules of Professional Conduct, including the principles of confidentiality, competence, UPL, and supervision. However, the recent guidelines and opinions from Florida and California can be summarized as—common sense.
But first, I start with a rewind to 20 years ago—in 2004, when Gmail was introduced to the public. At one time, attorneys were seemingly prohibited from using email in general because email was considered to not be secure. However, by the time Gmail came around, email was not prohibited, but there were best practices. See, e.g., ABA Formal Ethics Op. 99-413. Nevertheless, years later, Gmail was on the radar of ethics lawyers due to the scanning of e-mails to target advertisements. 20 years later, I see Gmail and GSuite used freely by many lawyers.
Similar to how the use of cloud technology and email has changed as technology (and the reception to the technology) has evolved, a lawyer’s use of AI should be considered in the same way. Until the adoption of AI technology becomes commonplace, lawyers have a duty to investigate how this technology may impact their practice, and their clients. As an example, a lawyer’s competence requirement would require them to understand how the technology works, and to ensure information is kept confidential. The level of review may continuously change according to the adoption of the technology, but also according to the client’s sophistication.
So what has the USPTO said about AI? In addition to guidance on inventorship, the USPTO issued a Memo on the Applicability of Existing Regulations Related to the use of AI. The memo highlights that existing rules, including the USPTO Rules of Professional Conduct, and specifically 37 CFR 11.18, are “adequate to address the challenges that the USPTO is likely to face” noting that more guidance is forthcoming. One area that I have frequently suggested is missing is the USPTO’s view of Rule 56 related to the use of LLMs and AI.
Until further guidance is issued by the USPTO, one of the best ways for lawyers and law firms to evaluate their use of AI is to reduce to writing their analysis of the platforms or tools they wish to use, including answering the following questions:
- Do you understand the technology?
- How did you vet the technology?
- Did you inform your client (and do you need informed consent)?
- Are you reviewing the work product before it is submitted to the client or USPTO?
- Are you maintaining confidences?
- Are you supervising the use of AI?
As the use of AI (and AI itself) evolves, lawyers should continue to monitor how the use of AI is adopted by their peers, and how new issues may arise.
Disclosure: I have advised a number of AI-driven patent drafting tools being used by firms across the country.