What Kids Probably Already Know About AI (and Get Wrong)

Transcript

Most kids can get a chatbot to do just about anything. Ask them how it works, though, and you'll hear the same few wrong answers.

"It looked it up." Unless there's a link, it probably didn't. It wrote the words most likely to come next, based on a huge pile of writing. That's how it can sound sure and still be wrong.

"I asked if it was sure, and it said yes." That isn't a check. Sometimes it says yes no matter what. Sometimes it caves and changes a right answer. Either way, all you learned is what it says when you push. If it matters, find a source you can open.

"It knows me. It remembers my dog's name." It stored that. Storing isn't caring. Ask them, has it ever noticed you were quiet today without you telling it?

None of these are dumb guesses. They're exactly what you'd think from using it every day. That's the catch. Using it doesn't teach you how it works.