Intro to Legal AI

An AI assistant is a chat-based tool you can talk to in plain English. Claude, ChatGPT, and Perplexity are examples. You can ask them to explain ideas, summarize text, draft language, compare options, make a plan, or ask useful follow-up questions.

A useful way to think about an assistant is: it is a very fast helper for reading, writing, organizing, and brainstorming. It is not a lawyer, judge, court clerk, or legal database.

Your job is to give the assistant a clear task, safe facts, and enough context, then decide what can be used, what needs checking, and what should be ignored.

Open your preferred AI assistant, such as Claude, ChatGPT, or Perplexity, in a new browser tab. Use an account or plan that your organization allows.

For this first exercise, any plan that lets you send a normal chat message is enough. You do not need paid legal-research features yet.

Paste the prompt below as-is, with no client facts. The goal is just to see how the conversation feels.

For ordinary writing or brainstorming, an assistant can often be useful on its own. Legal research is different because the answer has to rest on real authority you can inspect.

Sometimes an assistant answers from its own built-in knowledge instead of checking a legal source. You may hear people call that "model memory" or "model knowledge." In plain English, it means the assistant is relying on what it already seems to know, not on a citation, quote, or source text it just looked up.

That is risky for legal work. The assistant might summarize a rule too broadly, miss a jurisdiction limit, overlook later treatment, or sound confident about a case that does not support the point you need.

It is also a lot to ask Claude, ChatGPT, or Perplexity to find a group of long opinions, read each one closely, compare facts and posture, check treatment, and decide which authorities are actually useful, all without a legal source layer designed for that job.

When a legal question needs real authority, the assistant is more useful when it can work from legal source material you can inspect. That means citations, cases, statutes, rules, quoted language, source text, and treatment signals you can inspect, not just an answer that sounds polished but does not show its legal support.

The point is not to replace the assistant or replace your judgment. The point is to give the assistant better legal material to work from, then keep you close enough to check the result.

You will start with safety and setup, then move through the same path a careful researcher would use: frame the question, build a roadmap, find authority, ask follow-up questions, and check citations, quotes, and source support.

The reading lessons are public. In the hands-on exercises later in the course, you will learn how to use source-backed legal research tools to move from a first question to research you can inspect.