Phishmael & FriendsKids are meeting AI everywhere: homework helpers, chat apps, filters and games. In this episode, the Reef meets three AI tricks, and learns three simple rules.
Rule 1: AI can be wrong. Chatbots can sound completely sure and still be wrong. They learned from huge amounts of writing and predict likely answers, so important facts should be checked in a book, with a teacher, or on a trusted website. (In the episode, a chatbot claims clownfish can fly on Tuesdays. Bubbles, a clownfish, has never flown.)
Rule 2: never tell a chatbot your secrets. Some apps pretend to be friends to collect personal details. Children shouldn't share their full name, address, school, passwords or photos with chatbots or AI apps. And a good app, or a good friend, will never ask a child to keep secrets from their family.
Rule 3: seeing isn't believing. AI can create realistic photos and videos of things that never happened, including famous people (or whales!) giving away free stuff. Check the real account, and ask a grown-up.
AI isn't bad: it's a tool, and it can be brilliant for learning. The episode's rule sums it up: smart robots, smarter kids. Check before you trust.