In the article “USPTO Outlines Possible Pitfalls For Attorneys Using AI,” Knobbe Martens partner Mauricio Uribe provided insight into the U.S. Patent and Trademark Office’s latest guidance on the practitioner’s use of artificial intelligence (AI) in matters before the USPTO.
Uribe maintained that “everything revolves around the duty of candor” in the guidance, which reiterated the USPTO’s existing policies advising attorneys to review any work generated by AI tools in a patent filing before submitting to the office. The guidance is not formal rulemaking by the USPTO and does not appear to signify any new or significant policy change by the USPTO. Uribe believes that it serves primarily as a practical guide and cautionary publication to reinforce best practices and highlight potential pitfalls.
Referring to several specific examples in the guidance that advised attorneys to be cautious about using AI to generate research and case citations as part of Information Disclosure Submission (IDS) practice, Uribe said the policy essentially said that “a human should not sign that form without having looked at all those references.” Uribe also commented on a related policy in the guidance that attorneys must ensure wording generated by AI for patent filings does not include information not considered by inventors. Speaking to the nuance of this policy, he said, “Even if you are careful about what the human inventor created, the act of drafting could inject that issue.”
Offering an overarching perspective on the increasing use of AI tools in legal work, Uribe expressed that early adoption of such tools has caveats, stating his “general position is that perhaps it’s never best to be on the leading edge of these things.” However, highlighting the ubiquitous and rapidly evolving nature of AI use in professional services, he added, “I’d say the guidance is more precautionary, but if you haven’t evaluated tools like this yet, you soon will be.”
Read the full article here [subscription required].
