AIQ
AIQ
Part of the When AI Gets Confused lesson guide. Teaching a different grade? 🔧 Builder (8–10) · 💻 Hacker (11–14) · ⚡ Architect (15–18)
Gather the class close — this lesson works best sitting together, not spread out at desks. Start with a silly face or a mixed-up moment of your own: "This morning I put my shoe on the wrong foot and didn't notice until I stood up! Has that ever happened to you — you got a little confused and did something silly by accident?" Let two or three kids share (calling a teacher "Mommy" by accident is a classic — if nobody offers it, you can).
Then say: "Guess what — even A.I. gets confused sometimes! Today we're going to catch A.I. making silly mistakes." Open the app and read the hook line together, with big energy on the funny part:
"A.I. is smart, but sometimes it makes really silly mistakes! 🤪
Like thinking a banana is a toaster! Let's see what confuses A.I.!"
Pause on "a banana is a toaster" and let the class react — this should get a laugh. Ask: "Do you know the difference between a banana and a toaster?" (Yes!) "A.I. does too, most of the time — but today we'll see how it can still get mixed up."
Play a thirty-second warm-up game called "Mirror Mix-Up" to make the idea physical before you open the app: stand facing the class and say "I'm going to be a robot copying you, but sometimes I'll get confused and copy the wrong thing!" Do a few simple actions (touch your nose, wave, clap) and have the class copy you correctly each time — then deliberately "glitch" once, touching your ear instead of your nose right after you said "nose," and freeze with a puzzled face. Kids will laugh and correct you. Say: "See? Even a pretend robot can get confused! Real A.I. gets confused too, in its own way — let's go find out how." That's the whole warm-up: kids get confused sometimes, A.I. gets confused sometimes, and that's okay as long as we're careful and always double-check with a grown-up.
Walk through the app's three picture groups together as a class, reading each card out loud and reacting to the picture and emoji before moving on. Keep this fast, physical, and playful — Explorer mode has no failure language and no pressure, so treat every example as a fun "aha," never a test. Budget about a minute per card; the whole activity should feel more like a picture-book read-aloud than a lecture. If your device has a projector, showing the cards on the big screen works better than passing one tablet around; if not, walk the tablet or phone slowly past each row of desks so everyone gets a look at the picture before you move to the next card.
Group 1 — Seeing Mix-Ups 👁️. Show the Stop Sign card (🛑) first: "Some clever grown-ups put tiny stickers on a real stop sign, and it tricked the computer into reading it as a different sign! Just a few little stickers!" Hold up your hand like a stop sign and ask the class to shout "STOP!" — then say "you weren't tricked by stickers, were you?" Move to the Banana card (🍌): "Someone hid a tiny secret change inside a picture of a banana — so tiny that YOUR eyes can't even see it — and the computer thought it was a toaster!" Ask: "Can you see any difference in a regular banana picture?" (No.) "That's exactly the trick — you can't see it, but the computer got confused by it." Then the leopard-sofa card (🐆): "A computer once looked at a comfy, spotty couch and said 'that's a leopard!'" Ask the class to make their best surprised-leopard face. Finish with the silly-glasses card (😎): "There are special glasses that can trick a computer camera into thinking you're a totally different person! Like a superhero disguise, but for computers." Let a couple of kids pretend to put on invisible "disguise glasses."
Group 2 — Talking Mix-Ups 💬. Read the sarcasm card (😄) out loud in your most tired, sarcastic voice: "Ohhh great, ANOTHER Monday." Ask the kids: "Did I really mean I love Mondays?" (No!) "A.I. might think you did, because it just hears the words, not your grumpy voice." Next, the same-word card (🏦): "The word 'bank' can mean the place that keeps your piso money, OR the side of a river! A.I. sometimes mixes up which one you mean." Then the nonsense card (🤪): "A.I. can write a whole confident paragraph about 'why rocks make great pets' — sounding very sure of itself, even though it's silly!" Ask a volunteer to say one sentence about why a rock would make a great pet, in their most serious voice, for a laugh. Finish with made-up facts (📅): "Sometimes A.I. tells a story that sounds true and sounds sure of itself — but it's actually just made up, like a very confident guess."
Group 3 — Why This Happens 🔍. Keep this part very simple and go card by card: "A.I. only knows what it has seen before (📊) — like if you never saw a purple banana, you might not know what to call it either!" "A.I. doesn't really *know* things the way you do (🧩) — it's more like matching puzzle pieces, so it doesn't automatically know a car can't fly." "A.I. can't always tell what a picture *means* (🌍) — like it can't always tell if a toy is a real one or just a toy." "And grown-up engineers work hard to test A.I. and make it safer (🛡️) — but it's never going to be perfectly right 100% of the time, and that's exactly why we always check with a grown-up."
Pacing tip: with a wiggly group, you don't need every single card — the twelve pictures across the three groups are meant as a menu, not a checklist. If energy is dropping, pick your two or three favorite cards from each group and move briskly to the discussion questions instead of forcing all twelve. What matters for Explorer mode is the feeling ("computers get confused, so we double-check"), not memorizing every example.
Bring the class back into a circle or gather them on the rug for this part. There's no need to call on every child for every question — two or three answers per question, with lots of nodding and "ooh, good one!" from you, keeps the pace right for this age and keeps it feeling like a conversation rather than a quiz.
The in-app quiz runs right after the lesson, one question at a time, with big friendly buttons and no timer — there's no wrong-answer penalty in Explorer mode, only Pixel's gentle "let's look again" if a child taps something else. You can also ask these four questions out loud as a group before or after the app quiz, as a quick oral check.
Close with: "Today you became A.I. detectives! You learned that even smart computers get confused sometimes — just like we do — and that's exactly why we always check with a grown-up before we believe something completely. You did such a great job spotting the silly mix-ups!" If the class has stickers or stamps available, this is a nice moment for a "A.I. Detective" sticker, in keeping with Explorer mode's star-and-sticker rewards rather than any competitive prize.
Extension 1 — "Spot the Silly Mix-Up" drawing: Hand out paper and crayons and ask each child to draw one thing that might confuse an A.I., inspired by what they just learned — a banana that looks a little like a toaster, two animals that look almost the same, or a funny face making a joke. Give them three or four minutes. Then let volunteers hold up their drawing and explain in one sentence why it might be tricky for A.I. — "my drawing has a cat that looks like a dog because of the ears!" is exactly the level you're after. Collect the drawings on a "A.I. Gets Confused" bulletin board if you have wall space; it becomes a nice callback the next time the class talks about computers.
Extension 2 — "Confused Robot" role-play: If there's a few extra minutes, pick one child to be the "A.I." and give them a very simple, silly rule, like "say 'apple' every time you see something round." Show a series of objects or pictures (a ball, an orange, a book, a wheel) and let the "robot" apply the rule out loud, including on things that aren't actually apples. Ask the class afterward: "Was the robot being naughty, or just following its rule too simply?" This physically reinforces, without any new words, the core idea of the lesson — that A.I. follows patterns, not real understanding, and that's exactly where mix-ups come from.