AIQ
AIQ
Everything a teacher needs to deliver this lesson — pick your grade's script below once you've read the background.
"Your AI Rules" is the capstone of World 5 (Is AI Fair?). Every earlier lesson in the world — bias, privacy, deepfakes, environmental cost, and accountability — gets pulled together here into something the student writes themselves: a personal list of rules for how AI should behave. There is no new AI concept introduced in this lesson; the point is synthesis and ownership. By the end, a student should be able to:
The depth of "explain why it matters" scales hard across the four age bands: a 6-year-old is reasoning about fairness with a "would that be nice to my friend" test, while a 17-year-old is expected to weigh competing values and precedent. Use the age-specific script (linked below) — this page only covers what is common to all of them.
Because this is the final lesson of World 5, it is also a good checkpoint for you as the teacher. If a student can articulate their own version of these five objectives in their own words — not recite the app's phrasing back — that's a stronger signal that the whole world landed than any single quiz score would be, since the quiz questions can be answered correctly by pattern- matching without real understanding underneath.
You do not need an engineering background to teach this lesson well — it is mostly about values, not mechanics. Everything the students already learned about how AI actually works (in the earlier World 5 lessons) is being applied here, not re-taught. Your job in this lesson is to help students turn "AI can be unfair, AI can be wrong, AI can be misused" into "here is what I think should be done about that."
The lesson organizes its rules around four ideas, and it helps to have your own plain-language grip on each before you stand in front of the class:
Fairness means an AI system should not work better for some groups of people than others just because of who trained it or what data it learned from. A face-unlocking app that struggles with darker skin tones, or a hiring tool that quietly favors one gender, are both real, documented failures of fairness — not hypothetical ones. This lesson does not re-teach the mechanics of bias (that is lesson 5.1.1); it just asks students to state fairness as something they'd demand.
Transparency means a person should be able to find out that an AI made a decision about them, and get some kind of explanation of why. A simple example students will recognize: if a streaming app recommends a video, that's low-stakes and nobody needs an explanation. If a system decided whether a loan application, a college application, or a school placement was accepted, the person affected has a real claim to know why — and to challenge it if it seems wrong.
Safety in this lesson means testing an AI system thoroughly, especially before it is used somewhere a mistake could really hurt someone — a self-driving car, a medical diagnosis tool, a system that recommends prison sentences. The lesson's framing is simple: the higher the stakes, the more testing is owed before deployment.
Human control is the idea that a person should always be able to step in and override an AI's decision, particularly when the decision affects someone's safety or life. This is the principle behind why, for example, a fully autonomous weapon or a fully automated medical life-support decision worries ethicists even when the underlying AI is statistically accurate — the concern isn't the accuracy, it's the absence of a human who can say "wait, stop."
The lesson then extends past principles for AI-makers into a second idea worth spending real class time on: the student's own leverage as a user. Every app downloaded, every permission granted, and every piece of feedback given (a bad rating, a reported bug, a complaint that a filter is behaving unfairly) is a small vote about what AI companies build next. This is not exaggeration for the lesson's sake — user reports and public pressure are genuinely how several well-known bias and safety problems in commercial AI products got fixed. Framing this for students as real agency, not empty encouragement, is the difference between this landing as inspiring versus landing as a platitude.
Finally, the lesson insists that someone — a specific person, team, or company — is accountable for every AI decision, and that "the algorithm decided" is not an acceptable answer. This is a real and current debate in AI policy (it shows up again, in more legal depth, in lesson 5.2.2, "Who's Responsible?"), and this lesson's job is just to plant the idea that accountability cannot be automated away.
No special setup. Each student (or pair, for younger grades) needs a device with a browser to open AIQ and reach lesson 5.2.3 under World 5. Since this is a synthesis lesson, it lands best when students have already completed the rest of World 5 (bias, privacy, deepfakes, environment, accountability) — if your class hasn't, a two-minute verbal recap of "what has AI gotten wrong that we've talked about" before starting will give the "your rules" activity something concrete to stand on.
Have paper or a shared doc ready for the writing activity in the Main Activity section — the in-app lesson doesn't collect the student's own written rules, so that has to happen outside the app if you want a record of it. Nothing else needs to be printed, purchased, or set up in advance — this is one of the lower-prep lessons in the curriculum precisely because it's meant to be a discussion-and-synthesis session rather than a hands-on build.