AIQ
AIQ
Part of the AI Is Everywhere! lesson guide. Teaching a different grade? 🌈 Explorer (5–7) · 🔧 Builder (8–10) · ⚡ Architect (15–18)
Hacker mode drops the mascot and the softer framing entirely — no Pixel, no Axiom, just a direct address to the student. Open it like you'd open a real investigation rather than a story. Read the app's hook line directly:
"Here's something wild: you've probably interacted with AI dozens of times today without realizing it. Let's break down exactly where AI shows up in your daily life — and where it doesn't."
Push it further before opening the app: "I want you to name one piece of tech you used in the last hour that you're SURE is not AI, and defend it." Take two or three answers and press on the reasoning — "why are you sure? What would change your mind?" This sets up the actual analytical skill the lesson is teaching: not just labeling things AI/not-AI, but being able to justify the label with a rule.
This age group tends to have opinions already about AI, often from social media rather than firsthand experience, and some of those opinions will be overconfident in either direction ("AI can do literally anything now" or "that's not real AI, that's just an algorithm"). Don't shut those opinions down — use them as data. If a student dismisses something as "just an algorithm," that's actually a useful moment: ask them what they think the difference is between "an algorithm" and "AI," since the honest answer is that AI systems are themselves a category of algorithm, just one that's learned from data rather than been fully hand-specified.
Work through the lesson's three scenes, but frame the walkthrough around the rule the app states directly in its own age-band summary: "AI systems process large datasets to identify patterns and make predictions. Unlike simple programs that follow fixed rules, AI adapts its behavior based on new data." Write that sentence on the board before you start, and have students apply it item by item instead of just recalling app labels. The goal for this age group is to move past label-recall entirely — by the end of the activity, a student should be able to look at something the lesson never showed them and reason through whether it fits this rule, rather than needing to have seen the specific example before.
Morning 🌅: Smart Alarm (AI — a voice-assistant speech-recognition model), Shower (not AI — fixed plumbing), Face Unlock (AI — a facial-recognition model trained on many faces to distinguish yours), Light Switch (not AI — a simple circuit). Ask: "What's the input and the output for the face-unlock model?" (Input: a camera image. Output: a yes/no match against a stored face profile.) This input/output framing is worth introducing early and reusing for every item in the walkthrough — it's a habit that pays off well beyond this one lesson, since "what goes in, what comes out" is the first question worth asking about any system, AI or not.
School 🏫: Spell Check (AI — trained on huge volumes of real text to learn word patterns like "their" vs. "there"), Paper Book (not AI), Game Levels (AI — difficulty adapts based on tracked player performance), Scissors (not AI). Ask: "Spell check used to be simpler — a fixed dictionary lookup. What changed to make modern spell/grammar check AI rather than a lookup table?" (It now predicts likely intended words/phrasing from context, which needs a trained language model, not just a word list.) The game-difficulty example is worth a beat too: ask what data it's actually using — typically your recent win/loss rate, reaction time, or number of retries on a level — to distinguish it from a game that simply gets objectively harder on a fixed schedule regardless of how you're doing.
Evening 🌆: Song Picks and Video Ideas (both AI — recommendation systems built on listening/watching history at scale), Bicycle (not AI), GPS Maps (AI — the routing itself is deterministic math, but traffic prediction layered on top of it is a learned model). This last one is worth flagging explicitly: GPS positioning is not AI, but the traffic-aware routing built on top of GPS data is. That's a genuinely subtle distinction and a good one for this age band to sit with — it shows that "is this whole app AI?" is often the wrong question; the better question is "which specific piece of this app's behavior is AI, and which piece is precise engineering it's built on top of?"
Close the activity on the app's key distinction, stated as directly as it is in-app: "a calculator follows instructions, but a recommendation engine learns your preferences over time." Have students restate that distinction in their own words before moving to discussion.
Then run the in-app practice round (ten more items) and use it as a fast diagnostic rather than a repeat of the same explanation. The set includes Siri/Alexa, Photo Filters, Smart Thermostat, Netflix Picks, Auto-Translate, and Spam Filter as AI, against Broom, Paper Clip, Metal Key, and Ruler as non-AI. Spam Filter is worth a quick aside: it's a good example of a classification task specifically, sorting incoming messages into two categories based on patterns learned from huge amounts of previously labeled spam and non-spam email — a slightly different shape of problem than the recommendation systems (song picks, video ideas) discussed above, even though both fall under "AI."
Push students to disagree with each other on the first two, since neither has a single tidy answer. Reasonable class time here: 8–10 minutes, ideally with students first jotting a one-sentence answer individually before any group discusses out loud, so quieter students commit to a position before the loudest voice in the room sets the tone.
Close with: "The test you built today — fixed rules vs. learned-from-data behavior — is the same test researchers, journalists, and policymakers use when they argue about what counts as 'AI' in the news. You're now equipped to spot when a company is using the word loosely."
Also worth flagging as you close: this lesson has focused entirely on AI that quietly predicts or recommends things in the background. It hasn't touched generative AI — chatbots, image generators, tools that create new text or pictures rather than sorting or recommending existing ones. Students at this age have almost certainly encountered generative AI already and may be eager to talk about it; it's fine to acknowledge that directly ("we'll get to that soon — it's a different, newer kind of AI, and it deserves its own lesson") rather than either diving into it unprepared or pretending it doesn't exist.
Extension activity — Marketing Claim Audit: Bring in (or have students find, with guidance on reliable sources) 3–4 real product descriptions or ads that use the word "AI" or "smart" — a smart fridge, a "AI-powered" toothbrush, a fitness tracker, a smart speaker. In small groups, students apply today's rule to argue whether the claim is likely genuine AI (learns/adapts from data) or a marketing stretch (fixed automation dressed up as "smart"). Have each group present their verdict and reasoning to the class. This extends the 10–15 minute core lesson into a full period and builds media-literacy skills the curriculum revisits later (misinformation, generative AI).
A useful constraint for the audit: require each group to name specifically what data the claimed system would need to have been trained on, and to flag if the ad never actually explains that. A genuinely honest "smart" product description usually says something about what it learns from (your usage patterns, your voice, your schedule); an ad that just repeats "AI-powered" as a slogan without ever describing what it learns from is the clearest tell of marketing dressing up ordinary automation. This gives students a concrete checklist rather than a vague "does it feel legit" judgment call.