The short version

What this episode covers

AI hallucinations explained in plain English: what it means when AI states something totally made up with total confidence, and what that costs a real business.

AI doesn't just get things wrong. It gets them wrong with full confidence and no warning. That's a hallucination, and it's not a rare glitch, it's baked into how these tools work. In this episode of the AI Dictionary, two business owners break down what a hallucination actually is, how often it really happens (the number depends on one thing you control), and the exact habits that catch a made-up answer before it reaches a customer. No hype. No fear-mongering. Just the honest version.

What you'll take away

  • What an AI hallucination is and why it happens even in the best models
  • The error gap that matters: grounded AI (~2%) vs AI answering from memory (40%+)
  • The Air Canada chatbot case, and why "the bot said it" was not a defense in court
  • What happened to two law firms that filed AI-fabricated case citations
  • The five habits that prevent most hallucination damage before it starts
  • What RAG is, and why grounding is the biggest reliability upgrade you can make
The series

AI Dictionary

21 episodes so far, all free to watch.

See every episode →
Got a question about this one?

Bring it to the community, or bring it to us.

The free community has a live call every Thursday. If you'd rather talk about your own business, book a call and we'll tell you what's worth automating and what isn't.