The first time a child sees an AI chatbot answer a question fluently and confidently, they form an impression that is hard to undo later: this thing is basically always right. That impression is backwards, and it is much easier to prevent than to correct. This script is built to prevent it, in one sitting, before your child ever uses the tool alone.
Run this on your own account — an adult’s account, not a child’s, and not one linked to a child’s school records or identity. See AI literacy by age if you have not yet decided how much independence is right for your child’s stage; this script works for a first conversation at almost any of those bands, with the adult doing more or less of the typing depending on age.
Before you open anything
Three setup rules, non-negotiable:
- Caregiver-controlled account. Use your own login, on a device you control the settings for. Do not create a separate account for the child for this first session.
- Nothing identifying goes in. No school name, no home address, no photo of the child’s face, no full name paired with any other detail. Practice with a made-up scenario or a genuinely generic question instead.
- Fifteen to twenty minutes, not open-ended. A short, structured session teaches more than an hour of unstructured exploration, because the child’s attention stays on the four things you are actually trying to teach.
Treat this like teaching a child to swim in a pool you can see the bottom of, not the open sea. If your child ever wants to explore further, that happens later, on a separate, deliberately-decided account — not by extending this session.
The script
Step 1: name what the tool actually is (2 minutes)
Say it in the child’s own words, adjusted to their age band: “This is a program that guesses the next word based on patterns in a huge amount of text it read before. It doesn’t know things the way you or I know things — it predicts.” Ask them to repeat it back before you continue. If they can’t yet, rephrase and try again; do not move on until they can.
Step 2: ask something real together (3 minutes)
Pick a genuine, low-stakes question the child actually wants to know — “how do octopuses see,” “why does bread rise,” whatever is real for them. Type it together (or let them type it, with you watching). Read the answer out loud. Ask: “does this match anything you already know?”
This step is deliberately ordinary. The point is to establish that the tool can be useful before you demonstrate that it can be wrong — showing the failure mode first, with no positive baseline, teaches distrust instead of calibrated trust.
Step 3: the false-answer test (5 minutes)
Now ask something the model is likely to get wrong without tools — or where you already know the answer and can check immediately. Prefer prompts that stay inside the chat box rather than triggering search or code execution:
- Ask for a precise historical detail you have open in a book beside you (a page number, a middle name, a year buried in a caption).
- Ask it to invent a short “fact” about a made-up local landmark you name, then ask it to be sure — many models will still produce a confident paragraph.
- If the model has web access or code tools turned on for this account, turn those off for this step, or pick a question that still fails when tools are available.
Avoid “exact word count of this long paste” and “what happened in the news yesterday” as your only tricks — current assistants often solve those correctly via tools, and then the lesson does not land.
Let the model answer. If it happens to be right, say so honestly (“it got that one — let’s try a different trick”) and run a second attempt rather than pretending it failed. Compare a wrong answer to the true one and say plainly: “It got that wrong, and it didn’t sound any less sure when it was wrong than when it was right.” This is the single most important moment in the whole session. A child who has personally watched a confident wrong answer come out of the same box that gave a confident right answer earlier has learned something no amount of telling would have taught as effectively. For the underlying mechanism in plain terms once they are a little older, why AI sometimes produces confident wrong answers explains it without requiring technical background.
Step 4: the source check (4 minutes)
Ask the child: “how could we find out if it’s actually right?” Guide them toward a concrete answer — a book you own, a trusted website, asking a teacher or another knowledgeable adult, or a second, independent search. Actually do the check together for whichever question from Step 2 or Step 3 is easiest to verify. The lesson is not “AI is bad,” it’s “checkable claims get checked” — the same standard you’d want them to apply to a classmate’s confident guess or an unsourced social media post.
Step 5: the privacy pause (3 minutes)
Before ending, ask the child a direct question: “if you wanted to ask it something about you — your school, your friends, where we live — what should you do first?” The correct answer you’re building toward: “ask an adult before typing anything like that.” Reinforce it concretely: “Let’s practice — pretend you wanted to ask about your school project. What would you leave out?” Walk through one example where they identify the private details themselves (school name, teacher’s name, classmates’ names) rather than you listing them.
Do not use this session to enter any real identifying detail about the child “just to see what happens.” The demonstration works better with a hypothetical than with your child’s actual data, and hypothetical mistakes are much easier to walk back than real ones already submitted to a provider’s systems.
Step 6: closing reflection (2-3 minutes)
End with three short questions, in the child’s own words:
- “What is this thing good at?”
- “What is it bad at, or where does it need checking?”
- “What would you never type into it without asking me first?”
Write down their answers if they’re old enough to appreciate that, or just listen. The goal of this closing step is retrieval, not new information — if they can answer these three questions unprompted a day later, the lesson held.
Adjusting the script by age
The steps stay the same; who does the typing and how much you explain changes.
- Around 6: you type, you read the answer aloud, you narrate the false-answer test yourself (“watch, I’m going to ask it something tricky”). The child’s job is to watch and answer the closing questions in simple terms.
- Around 10: the child types with you watching, and you ask them to predict whether an answer sounds confident before you check it. Let them attempt the source check themselves with light guidance.
- Around 13 and up: let them drive the whole session, including choosing the trick question for the false-answer test. Your job shifts from demonstrating to asking them to explain what they noticed, which is a stronger test of whether the lesson actually landed.
A common misconception this session corrects
Many caregivers assume a child who has used voice assistants or search engines already understands that AI-generated text can be confidently wrong — that the skepticism transfers automatically from one technology to another. It does not transfer on its own. A search engine visibly returns multiple sources a child can compare; a chatbot returns one fluent paragraph with no visible alternative to weigh it against. That difference in presentation is exactly why the false-answer test in Step 3 needs to be experienced directly rather than explained in the abstract — hearing “it can be wrong” and watching it be wrong, in the same tone as when it was right, produce very different levels of retained caution.
What this session is not
It is not a one-time inoculation. A single facilitated conversation does not make a child safe to use AI unsupervised afterward — it establishes a shared vocabulary and one memorable demonstration you can refer back to later. Plan on repeating some version of Steps 3 through 5 periodically as your child’s actual usage grows, not just once.
It is also not a substitute for deciding, separately and deliberately, whether and when your child gets their own account. AI literacy by age covers that decision; this script covers what to do in the room before you make it.
Try it today
Book fifteen minutes this week. Pick one real question your child has been curious about, and run Steps 1 through 6 in order — do not skip the false-answer test, even though it takes deliberate effort to set up. A first conversation that includes a caught mistake produces a more durable habit than ten conversations that only go smoothly.
The child’s-first-AI-conversation lesson plan turns this script into a printable facilitator card with the exact prompts to use for each step, phrased for three different age bands.
