AIQ
AIQ
Part of the Learning from Mistakes lesson guide. Teaching a different grade? 🔧 Builder (8–10) · 💻 Hacker (11–14) · ⚡ Architect (15–18)
Ask the class: "Who here knows how to ride a bike?" Let hands go up, then ask one volunteer: "Did you get on the bike and ride perfectly the very first time?" They'll almost always say no.
"Remember learning to ride a bike? You fell, got up, tried again! 🚲 A computer helper called A.I. can learn the very same way — by trying, and failing, and trying again!"
Say it warmly, like you're letting them in on a secret: "Guess what — even smart computer helpers fall down sometimes, just like you did on your bike! And just like you, they get back up and try again, and again, until they get really good." Open the lesson on the app and show the four hook pictures — 🎮 Try, ❌ Fail, ✅ Succeed, 🏆 Master — and read each word out loud together. Point out that this is exactly the order things happened for their bike story too.
You can make the warm-up a little more physical for wiggly Explorer-age listeners: ask everyone to stand up and act out the four words with their bodies as you say them — a big excited "Try!" pose, a slumped "Fail" pose, a proud "Succeed!" pose with arms up, and a "Master!" victory jump. Do it once together, then say the four words out of order and see if they can switch poses fast. This gives the shape of the lesson — try, fail, succeed, master — a feeling in their bodies before you ever open the app, which matters more at this age than the words themselves.
The app walks through three little sets of pictures. Go through each one together, reading the fact out loud in simple words and pausing for reactions before moving to the next.
Show the Game AI picture. Say: "Imagine a computer playing a game — not once, but a THOUSAND times! Every time it plays, it tries something a little different, like moving left instead of right." Ask: "If it tried moving left last time and that didn't work so well, what do you think it should try next time?" Let a few kids guess — any answer that isn't "the exact same thing again" is the right instinct, and you can say so warmly even if their reasoning isn't precise yet. Show the Reward picture: "When it does something good, it gets a shiny star ⭐ — that's called a reward! When it does something not-so-good, no star. So next time, it remembers to do the good thing more!" Make this concrete with a game they know, like a simple matching or maze game on a tablet: "Imagine the game gives a little sparkle whenever you find a matching pair, and nothing when you don't. After enough tries, you'd start remembering which pairs go together — that's the reward doing its job!" Show Repeat: "It might play the SAME game a million times! That sounds like a lot — because it is! But every single time teaches it something new." Ask: "Could you play the same little game a MILLION times? How long do you think that would take you?" (Kids will usually say "forever" or "a hundred years" — a fun moment to point out that a computer can do it much, much faster than a person.) Show Improve: "At first it's just guessing. But after lots and lots of tries, it starts winning on purpose!"
Show AlphaGo: "There's a real computer program that learned to play a very hard board game called Go — by playing it against itself millions and millions of times! It got SO good, it beat one of the best human players in the whole world." Show Robot Walking: "Robots learn to walk kind of like a baby does — they fall down again and again in a pretend computer world, and slowly figure out how to balance, before they ever try walking for real." Ask: "Why do you think the robot practices in a PRETEND world first, instead of falling down for real right away?" (So it doesn't break itself while it's still learning — the same reason you might first learn to ride a bike with training wheels or on grass instead of a busy street.) Show Self-Driving: "A self-driving car practices driving smoothly and safely — good driving earns a reward, bumpy or unsafe driving doesn't, so it learns to drive more carefully over time." If any students have ridden in a jeepney, tricycle, or grab car, ask them what "smooth driving" feels like compared to a bumpy ride — that's the exact feeling the reward is trying to encourage.
Keep this part light — the words are big, but the idea is simple. "The computer that's learning is called the AI helper. The game or the road it's learning in is called its playground. Every time it does something, the playground gives it a star for good, or nothing for not-so-good — that's the reward. After lots of tries, the AI helper builds its own game plan for what to do next time." You don't need students to remember the words "agent," "environment," or "policy" at this age — "AI helper," "playground," and "game plan" cover the same idea. If a student asks what happens if the AI helper never gets a star, reassure them: "It just keeps trying different things until something finally works — the same way you kept pedaling and wobbling until one day you just... didn't fall!"
Close by reading the app's own summary together: "AI can learn by trying things over and over! Good tries get rewards, bad tries get penalties. Just like learning a game — practice makes better!" Ask the class to repeat the last three words back to you: "Practice makes better!"
Before moving to discussion, do a quick "thumbs check": say each of the three set titles again — Try, Try Again; Real Examples; How It Works — and ask students to give a thumbs-up if they remember what happened in that part, or a sideways thumb if they'd like you to say it one more time. This only takes a minute and tells you, before the quiz, whether any part needs a quick re-cap while the app is still open in front of them.
For this age band, there is no wrong answer here — the goal of the discussion is simply to get students naming their own real experiences of trying, failing, and trying again, and noticing that grown-ups (and computers) go through the same thing they do.
Close with: "So remember — falling off your bike wasn't a bad thing. It was YOU learning, just like the AI helper we talked about today. Trying, and not giving up, is how anybody — or anything — gets good at something."
Extension activity — Hot and Cold (10 minutes): Hide a small object (an eraser, a small toy) somewhere in the room while one student closes their eyes. Have the class guide them with only "warmer" (getting closer — that's their reward!) and "colder" (getting farther — no reward) as they try different directions. Let two or three students take a turn. Afterward, connect it directly back to the lesson: "You didn't know where the object was at first — you just tried a direction, listened for 'warmer' or 'colder,' and tried again. That's exactly how the AI in today's lesson learns!"
If time allows, add a quick drawing activity to close: give each student a small piece of paper folded in half, and have them draw themselves "trying and failing" at something on one side (falling off a bike, dropping a ball, a wobbly tower of blocks, missing a basket) and "succeeding" at the very same thing on the other side. Walk around and ask a few students to describe their two drawings in one sentence each. This gives the abstract idea of the lesson — try, fail, try again, succeed — a picture they made themselves, which tends to stick with 5–7 year olds far longer than the vocabulary does.
If you have extra time, a second option — Star Chart Game (10 minutes): Draw a simple three-step path on the board (three squares in a row leading to a star). Give a volunteer three tries to get a beanbag or crumpled paper ball into a small basket a short distance away. After each try, draw a star above the square if they scored, or leave it blank if they missed, and ask the class: "Do you think they'll aim the same way next time, or try to change something?" This makes the idea of a "reward" (the star) and "trying something a little different next time" visible and countable, which is exactly what the AI in the lesson is doing — just imagine that same chart with a million squares instead of three.