You’re up against a deadline. You run to ChatGPT. You tell yourself the privacy toggle will protect you and your clients confidential information. Guess what: it may not, at least in ways consistent with the ethical rules.

Lawyers and legal professionals may have gotten a little too lax about putting confidential client information into public-facing AI tools in part due to a  false sense of security that comes from toggling off that training switch. 

Here’s why the privacy settings may not meet the obligations under Model Rule 1.6, why privilege waiver is a threat, and some bedrock rules every lawyer should be thinking before hitting send on any prompt.

As the old Hill Street Blues captain used to say: let’s be careful out there..

My post for Above the Law.