Be Cautious When Using AI for Client Meetings
7.30.2025
Artificial intelligence tools can help during interactions with clients, but attorneys must make sure the programs used are private and secure.
The New York State Bar Association recently hosted a webinar delving into the promise and pitfalls of using AI tools for client meetings. More than 430 people attended.
The speakers were:
- Alexander Paykin, principal of the Law Office Of Alexander Paykin and chair of the New York State Bar Association’s Committee on Technology and the Legal Profession.
- Hilary Bowman, founder and CEO of Querious.
The speakers said there is a difference between general AI tools and those specifically built for the legal profession. While many general-purpose programs can take notes and summarize meetings, Bowman said that some tools are able to display relevant statues and bring up legal questions in real time.
“The goal is to enhance the in-meeting experience and be able to provide sharper advice during a conversation,” Bowman said.
However, attorneys should protect the privacy and security of their clients as well as the information they provide to an AI tool. Clients should be able to consent to the use of the technology and be informed of its purpose.
In addition, a tool must not feed private information to other AI tools or learning models. “Please don’t use the free version of any product,” said Paykin. “The terms of service generally make you the product.”
An attorney must also consider how long they would want to keep the information and notes generated by an AI tool.
Attendees asked many questions, including if the AI’s notes could be used in discovery.
“I think your cloud storage, your email, your AI chatbot assistant are all your tools,” said Paykin. “Like the data inside your computer or in your cloud folder is not discoverable and is generally considered protected, privileged, etcetera. To the extent your AI is working with you – on the case, on meetings with clients, on evaluating the data – that is your tool and whatever it generates is its work product.”
However, when another attendee asked if a client suing their attorney for legal malpractice could demand access to an AI system to see which suggested questions and advice it provided, Paykin answered that it could happen.
Bowman presented a list of questions that attorneys can use to evaluate an AI tool, including how the tool works, who can access the information, and if the information provided can be verified. This is a big concern given that AI tools can generate false or misleading information, otherwise known as hallucinations.
“I had a colleague many years ago who said something like ‘I’m not a techie, I’m not a tech lawyer,’” said Bowman. “That comment stuck with me. I don’t think any of us can take that approach today. Even if you’ve never taken a computer science class or you don’t have any background in technology, I think we all need to start asking questions.”
The program was sponsored by the association’s Committee on Technology and the Legal Profession, Committee on Law Practice Management, and Committee on Artificial Intelligence and Emerging Technologies. View the full seminar here.
