Risks in AI for Legal Work

I am often getting AI written notes and letters from clients, and maybe even other attorneys, and a recent article I came across gave me pause. The headline reads: “AI Docs Sent By Exec To Attys Not Privileged, Judge Says” LINK: https://www.law360.com/pulse/articles/2440082/

WOAH! So here’s the deal – communications between you and your attorney are generally privileged, which means that the other side cannot compel them in discovery. However, AI tools are not lawyers. So if you start asking an AI bot questions, not only can they seek that information from YOU, but they can subpoena the AI company as well. Most good lawyers would be able to tell you this, but it’s the second part that is scarier for litigants: what about using AI to draft documents that you then give to your attorney for the purpose of litigation?

According to the Court in this case, sending the documents to the lawyer did not suddenly convert them to privileged. Uh oh. So the bottom line for clients/potential clients is this: DO NOT FEED DETAILS YOU DON’T WANT THE OTHER SIDE TO KNOW INTO AI.

And what about lawyers? I have not used AI much, but am starting to dabble, and know lots of lawyers who use it frequently. While there could be a valid work-product claim, the standard terms of service are clear that there is no expectation of privacy. Uploading confidential client data comes with significant risks, not the least of which is the unethical dissemination of confidential client data.

Anyway, something to think about for all of you who want to treat AI like it’s your own personal lawyer.