Practical AI training for everyone
Module 1
6 min read
When you type a question into a search engine like Google, it goes and finds a real webpage with the answer on it. AI chatbots, like ChatGPT, do something completely different. They don't go find an answer. They guess, one word at a time, what a good answer would look like.
That's the big idea for this whole module: AI is a guessing machine, not a looking-up machine. A really, really good guessing machine — but a guesser, not a finder.
Imagine you read millions and millions of books, websites, and conversations — way more than any person could read in a hundred lifetimes. After reading all of that, you'd start to notice patterns: which words usually come after other words, how people usually explain things, how a story usually ends.
That's basically what happened to the AI, except a computer did the reading. It doesn't remember the books like a library card catalog. It just learned the patterns — so now when you ask it something, it predicts, word by word, what a good answer sounds like.
Try this comparison
Ask a parent: 'If I ask you what 9 x 7 is, you know it — you're not guessing.' Then: 'If I ask an AI to finish the sentence "Once upon a time, there was a..." it's guessing the next likely word, over and over, until it has a whole story.' Math facts are known. Stories (and a lot of AI answers) are guessed.
Try it now
Together, ask a chatbot: 'What is AI, in one sentence a 10-year-old would understand?' Read the answer out loud. Does it match what this module just said? Talk about anything it left out.
A few quick questions. You'll see an explanation after each one.
Question 1 of 4
What's the biggest difference between a search engine and an AI chatbot?