Non-Use of Generative A.I.

If I represent you, I promise I will not use generative artificial intelligence tools in any aspect of my work. I will not use A.I. to write emails, brainstorm, generate images, crunch data, review documents, and certainly not to draft documents to file in court. I do not believe the A.I. drafting tools widely available at this time, specifically large language models, currently have any place in the practice of law. If you're paying me, you get me. My work, my effort. Not some glorified suggested-text word generator's haphazard output. The dart-throwing robot is simply not helpful for making legal arguments, and may in fact be dangerous.

​

I am not an A.I. denialist. There are certain applications, for example document review, where A.I. tools such as large language models show great promise. Other attorneys I have spoken to have told me that there are currently LLM-aided doc review products that are highly effective. However, the list of lawyers who have used A.I. to disastrous effects is long and growing. These lawyers are not just being fined and disciplined; they're hurting their clients' cases. I believe this conduct is unethical, immoral, and flat-out bad, lazy lawyering.

 

I think it is possible that, someday, A.I. writing tools will be useful and reliable enough for regular use in the practice of law, with proper ethical and practical safeguards. But that day is quite clearly not today. Until such time, I refuse to put my clients' best interests at risk. This promise is so important to me that I pledge it in every single representation agreement I offer to a client.