Unless attorneys opt out, client information that they include in the prompt they feed into AI tools could later be used to train those same tools and published, leaving the attorneys in violation of the New Jersey Rules of Professional Conduct, according to a Tuesday virtual panel discussion organized by the New Jersey State Bar Association.
"It's sexier and more interesting to hear about these gaffs and these problems" with fake citations, according to one of the panelists, Carol Johnston of the state Supreme Court clerk's office, but the risk of accidentally exposing a client's name, other identifying information or details about their case is "the big takeaway from the ethics point of view of the peril of using AI."
"These programs will take your written submissions, your prompts, and use them for their own learning and spit them out elsewhere," Johnston said.
Johnston is a staff member and counsel to the state Supreme Court's Advisory Committee on Professional Ethics, the Committee on Attorney Advertising and the Committee on the Unauthorized Practice of Law, and also answers the New Jersey attorney ethics research assistance hotline.
The New Jersey Supreme Court has been wading into the debate on how AI can ethically be incorporated into the practice of law with a committee formed in September 2023, guidelines issued in January, and an anonymous survey issued to attorneys in March about how they're using AI.
Against that backdrop, the state bar convened a panel featuring Johnston and others on Tuesday to discuss the pitfalls of using AI in legal work.
One of the panelists, Chad Moore, sat on the Supreme Court AI committee that helped draft those guidelines and spoke to the prevalence that AI has already reached in the legal industry.
"You can't just assume anymore that people aren't using it," Moore said. "It's very widespread."
Moore is managing partner of Hoagland Longo Moran Dunst & Doukas LLP and also chairs the firm's AI practice group and technology committee.
Moore said he had used some generative AI tools and found them to be useful for getting a start, but they still have a long way to go. But even at this early stage there are numerous pitfalls where they intersect with New Jersey's Rules of Professional Conduct.
Moore highlighted, for instance, rules governing what attorneys must tell their clients as one such area. He said attorneys should update their retainer agreements to promise to tell clients if they use AI in working on their case.
"I think a client now is going to have to be informed if we are using AI in their matters," Moore said. "I think they deserve to be notified and told."
Glenn Reiser, an attorney ethics defense specialist at Shapiro Croland Reiser Apfel & Di Iorio LLP, said the fast-paced nature of the legal profession puts pressure on attorneys to use the fastest means available of working, but he cautioned that with AI the risk of violating the RPCs is such that attorneys should prioritize diligence.
"As fast as our clients want us to do things, I think it's important that we slow the process down and do things at our speed, so that we don't expose ourselves to problems like this," Reiser said.
AI presents complex challenges on several fronts, the panelists agreed. But as Reiser noted, it is not enough for attorneys to attempt to avoid the technology entirely.
"We have to have a general understanding of how this technology works," Reiser said. "We can't just claim ignorance."
--Editing by Robert Rudinger.
For a reprint of this article, please contact reprints@law360.com.