AIQ
AIQ
Everything a teacher needs to deliver this lesson — pick your grade's script below once you've read the background.
"AI Is Everywhere!" is the very first lesson in AIQ, so its job is narrow on purpose: give students a working eye for AI before anything else gets built on top of it. By the end of the lesson, a student should be able to:
None of this requires memorizing terminology. A student who can correctly sort ten random household items into "AI" and "not AI," and give one sentence of reasoning for each, has met the bar for this lesson — everything more technical than that (how the learning actually happens, what a dataset looks like, what "training" means step by step) is deliberately deferred to later lessons in World 1 and beyond. This lesson is a foundation stone, not a survey of the field.
You do not need a computer science background to teach this lesson well — the app does the heavy lifting with visuals and examples. What you need is a clear, correct mental model so you can answer the question every class eventually asks: "Wait, so is my calculator AI?"
The simplest working definition, and the one this lesson uses without naming it formally, is: AI is software that improves at a task by learning from data and examples, instead of only following rules a programmer wrote in advance. A traditional program is a fixed recipe — if the input matches condition A, do B. It never gets better at the task on its own; a person has to go back and rewrite the rules. An AI system, by contrast, is shown many examples (millions of photos of faces, millions of song-listening histories, millions of typed sentences) and works out its own internal rules for recognizing patterns in new, unseen examples it wasn't explicitly told about.
This is why the lesson's running test — "does it learn from data, or does it just do the one fixed thing?" — is genuinely the right test, not a simplification that will need to be walked back later. A light switch, a bicycle, and a pair of scissors do exactly what they were mechanically built to do, every time, with zero exceptions or improvement. A phone's face-unlock, a music app's recommendation list, and a maps app's traffic prediction all behave differently for different people and get more accurate the more they're used, because they were built on data and keep learning from it (or at least were trained on huge amounts of it). That's the dividing line the lesson's morning/school/evening scenes are built around, item by item.
It's also worth being upfront, even with younger students, about what today's AI can't do. The lesson explicitly ends on this point in every age band: AI finds patterns in data, but it does not feel emotions, form real friendships, or "understand" the way a person does, even when its output looks convincingly human (a chatbot's kind-sounding reply, a music app's "picked just for you" playlist). This isn't a philosophical aside — it's a load-bearing idea students will need throughout the rest of the curriculum, especially when the app later covers generative AI and misinformation.
One more thing to have ready: this lesson is deliberately about breadth, not depth. It does not explain how a neural network works, what training data actually looks like, or how a model is built — that's the job of later World 1 lessons (1.1.2 "Smart vs. Intelligent," and lessons further in on how computers see, hear, and read) and later worlds entirely. Here, the only goal is recognition: can the student point at something in their life and correctly say "AI" or "not AI," and say roughly why.
A useful way to think about your own role in this lesson: you are not teaching students a fact to memorize, you are installing a habit of noticing. The app supplies the examples and the correct labels; your job in the classroom is to keep asking "why do you think that?" after every answer, right or wrong, so the reasoning — not just the label — is what sticks. A student who can correctly say "Netflix recommendations are AI" but can't say why is at the same starting point as a student who has never opened the app; a student who can explain the "learns from data" test in their own words has actually learned something transferable.
Finally, a note on scope across the four age bands you'll see in the delivery scripts. The underlying content — the same three time-of-day scenes, the same list of AI and non-AI examples — is identical for every student in every grade. What changes by age is vocabulary, depth of the "why," and how far the explanation goes past the basic label. A 6-year-old and a 17-year-old both need to walk away knowing a phone's face unlock is AI and a light switch isn't; only the older student also needs to know that's called a computer-vision classification task. Don't feel you need to withhold correct-but-simple language from older students, or push technical vocabulary onto younger ones — each script is calibrated to what's actually useful at that age, not what sounds more impressive.
This is a low-prep lesson by design — it's meant to be usable in a computer lab, a one-device classroom, or at home with a single shared phone.