A Legal Research Prompting Guide and Generative AI System Comparison Exercise

I’m sharing a guide and exercise I’ve developed for my legal research courses. This Google spreadsheet provides instructions on crafting AI prompts for legal research and includes a practical exercise for comparing different AI systems. It’s designed to help develop skills in leveraging AI for legal research. Feel free to copy it to adapt it to your own purposes. (Note: The images were blurry unless I sort of chopped them off, so sorry about that!)

The spreadsheet consists of three different parts:

Prompt Formulation Guide: This section breaks down the anatomy of an effective legal research prompt. It introduces the RICE framework:

  • R: Role (assigning a role to the AI)
  • I: Instructions (specific tasks for the AI)
  • C: Context (providing necessary background information)
  • E: Expectations (clarifying desired outcomes)

Sample Prompts: The spreadsheet includes several examples of prompts for various legal research scenarios which can serve as templates.

AI System Comparison Exercises: These sections provide a framework for students to test their prompts across different AI systems like Lexis, ChatGPT, and Claude, allowing for a comparative analysis of their effectiveness.

Feel free to copy it to adapt it to your own purposes, and let me know if you have any suggestions for improvements!

4 thoughts on “A Legal Research Prompting Guide and Generative AI System Comparison Exercise

  1. Pingback: Law Librarians Leading the Way on AI | New Ideas Legal Tech

  2. I absolutely love this, but the problem is in the context portion client information is being disclosed. If ChatGPT4o or Perplexity.ai Pro are being used the licenses (even for the subscription edition) give those platforms the right to use inputs and outputs including private information.
    P.S. I’d love to help write for this blog.

  3. Pingback: Announcing the AI Law Librarians Prompt Library - AI Law Librarians

Leave a Reply

Your email address will not be published. Required fields are marked *