AI Practice Docket – Episode 2: Understanding Generative AI and Its Limits

This episode explains why generative AI systems—while powerful—must be used with caution in legal practice. Unlike research databases, these tools generate content by predicting patterns, not retrieving facts. That means the same prompt can produce different results, and outputs may include errors or “hallucinations.” We discuss why human verification is essential, where generative AI can be helpful—like drafting and brainstorming—and where it falls short, such as legal research or citation verification. Think of it as a brilliant but unlicensed intern: useful, but never unsupervised.