Free family workout · 20 minutes · ages 7–17
The Second-Ask Drill
Most kids ask an AI one question and stop at whatever comes back. Fluent people don't. This workout builds the skill that separates the two: treating the first answer as a rough draft you direct — with a second ask, and a third — not a verdict you accept.
What you'll need
- Pen and paper for your child
- One AI chat tool on your device, used together — your child never needs their own account
- One real, open-ended question your child genuinely cares about (how to train for a race, plan a small party on a budget, understand something confusing from school, design a game, settle which phone or bike is actually worth it)
Why this matters (read this part to yourself first)
"AI fluency" gets talked about as if it means knowing how the technology works. For a child, it means something simpler and more useful: knowing how to direct the tool instead of being handed a result by it. The beginner and the fluent user type the same first question — the difference is what happens next. The beginner reads the first answer and stops. The fluent user reads it as a starting point and steers: adds what the machine didn't know, asks it to try a different way, asks where it might be wrong. That back-and-forth is the whole skill, and it's the opposite of passively accepting output. Today your child practises being the one giving directions across several turns — staying the director, not the audience.
Step 1 — Ask the first question (3 min)
Have your child type their real question into the AI, in their own words, and read the answer out loud together. Then say the ground rule before anyone reacts to it: "That's the first answer, not the final one. A first answer is where a good thinker starts, not where they stop. Our job now is to make it better." Resist the urge — and help them resist it — to just take what's on the screen and move on. The first answer is raw material.
Step 2 — Add what it couldn't know (4 min)
The AI answered a stranger. Your child knows things it doesn't. Ask them: "What did the answer assume about you that isn't true, or leave out that matters?" Maybe it suggested a budget that's too high, an age that's off, a plan that ignores something specific to your family. Have them write one or two of those facts down, then feed them back in: "Actually, I'm 11 and I have $20, not $200 — redo it for that." Watch how much better the second answer gets when they supply the context. That's the lesson landing: the quality of the answer depends on the quality of the direction.
Step 3 — Make it try a different way (4 min)
Now practise steering on purpose. Have your child pick one of these follow-ups and use it: "Give me a simpler version." · "Show me a completely different approach." · "Explain it like I'm the age I actually am." · "Give me three options instead of one." The point isn't which they choose — it's the felt discovery that they can change the shape of what comes back. A fluent user isn't stuck with the first format; they ask for the one they need. Let them run two or three of these and notice how the tool bends to a clear instruction.
Step 4 — Ask it where it's unsure (4 min)
The most powerful ask a fluent person makes is the one beginners never think of: "What might be wrong or missing in what you just told me? What should I double-check myself?" Have your child ask it, and read the reply together. Sometimes the AI will flag a real weak spot; sometimes it'll be vague. Either way your child learns the habit that keeps them in charge — you interrogate the tool, you don't just trust it. End by asking your child: "Now that we've gone back and forth, what's your own answer — what will you actually do?" The decision is theirs; the AI was a consultant they directed, not a boss they obeyed.
Step 5 — Bank the win (3 min)
Have your child finish this sentence on the page: "The first answer said … but after I steered it, my answer is …" That gap — between what one lazy question returns and what a few good follow-ups produce — is exactly the skill this whole course is about. Keep the page. Over a few weeks it becomes proof your child can point to: I don't just take what the machine says first; I direct it until it's actually useful.
The everyday version
Next time your child shows you something an AI gave them — an answer, an essay opening, a plan — your family's one-line habit is: "That's just the first answer. What's your second ask?" It takes five seconds and it retrains the reflex from accept to steer. Fluency isn't a trick you teach once; it's this small question, asked enough times that directing the tool becomes automatic.
Safety rules we use in every workout
- A parent is present whenever AI is used. AI tools have their own age requirements — the parent's account, the parent's device.
- Never enter your child's name, school, photos, or any personal details into an AI tool. When Step 2 asks your child to add context, keep it to the task (budget, age range, goal) — never identifying information.
- The final decision always belongs to your child, out loud or on paper — the AI advises, it never decides.
Liked this? There are 50+ more.
This is one workout from the AI-fluency week of the 10-week Human Edge System. The full course gives you a parent lesson, conversation script, workout, and reflection prompt for every week — covering agency, critical thinking, AI fluency, communication, resilience, money sense, and purpose, plus a final portfolio your child builds and keeps.