It’s likely that we all received by now from clients various AI generated outputs we were asked to address in one way or the other. For example, a client may use an AI tool, such as ChatGPT or Grok, to generate a draft of a legal document that may be needed in the client’s legal matter, and then asks the lawyer to review, provide feedback or even use. Based on my current understanding of the AI tools, like the ones mentioned above, and some real-life experiences, here are some preliminary observations I came up with, which seems that us lawyers, and clients alike, can only overlook at our own peril:


·     AI ‘HALLUCINATION’. Clients may not be fully aware of the potential for AI ‘hallucination,’ which may cause their overreliance on the AI output;

·     GIGO PRINCIPLE. The AI tools appear to follow the GIGO principle – Garbage In, Garbage Out - meaning that the quality of the prompts provided to the AI tool directly impacts the quality of the AI output; the prompt development appears to be a combination of art and science in itself, which resulted in the birth of new careers, such as ‘prompt engineers’;

·     EXPERT INPUT. Related to the GIGO principle is the conclusion that crafting great AI prompts require expertise in the respective subject matter, the law in this case; i.e., you really have to know what to ask the AI tool to do;

·     EXPERT OUTPUT ASSESSMENT. Evaluating the AI output unequivocally requires expertise. How do you evaluate the impact of the tone, the terms used, and points made in a legal document, e.g., a cease-and-desist letter, on the receiver of that letter, if you are not a lawyer, let alone a lawyer experienced in the respective area of the law? You simply can’t! You could, however, cause of a lot of damage, such as killing the potential for a desired outcome, if you send out that AI generated letter!


Does anyone see any other AI risks?


To conclude, it seems we lawyers are challenged here in many ways, including to finding a way to respectfully and effectively educate our clients about these AI-related risks. How do you do it?


#lawyer

#AI

# AI Lawyer