On a weekly basis, I am seeing "legal briefs" from clients seeking a consult that are clearly AI generated. They give a summary of their case, include specific and identifying information, summarize confidential documents and proffer "AI's legal opinion". Almost universally they are done with open AI tools. While AI can be incredibly useful when you have a legal problem, there are significant risks in using it to “brief” your lawyer:

• Privilege, confidentiality & privacy: What you tell AI may not (and likely doesn't) have the protections of what you tell your lawyer. By entering identifying and personal information, you may be breaching confidentiality provisions.
• Confident mistakes: AI can miss evidence, invent connections and misunderstand the law.
• More work, not less: Your lawyer still has to verify the summary against the actual evidence.
• False confidence: A polished answer can make a complicated legal problem look deceptively simple.

‍

AI can be useful to organize your thoughts, understand concepts and come up with better questions, but the best prompt is still: “Can we talk?