Most households are running one of two policies right now. Either “we don’t talk about it,” or a blanket ban that everyone quietly ignores. Neither one teaches a kid anything about the tool they are going to use for the rest of their working life.

Here is a framing that works better, because it draws the line where the line actually belongs.

The line is learning, not tools

Forget “is using AI cheating.” Ask instead: did the learning still happen?

A calculator does not prevent a kid from understanding math — unless they never learned what the operations mean. Same principle here, and it sorts almost every situation cleanly into three buckets.

  • AI does the work. It writes the essay, solves the problems, produces the thing that gets submitted. The learning did not happen. This is the line, and it is worth naming plainly.
  • AI explains the work. “I don’t understand why this step happens — explain it three different ways.” “Quiz me on this chapter.” “What am I missing in my argument?” The learning absolutely happened, often better than it would have alone. A patient tutor who never sighs is a real gift to a struggling kid.
  • AI checks the work. They wrote it, then asked for feedback. This is fine and genuinely valuable — with one caveat below.

The test that settles arguments

Ask them to explain it to you without the screen.

If they can walk you through their reasoning, defend a choice, and answer “why did you do it that way” — the learning happened, whatever tools were involved. If they cannot explain their own submitted work, it is not theirs, and they know it. This test is fast, hard to game, and it moves the conversation off enforcement and onto understanding, which is where you wanted it.

The caveat on feedback

AI feedback is confident and often generic. It will happily rewrite a kid's distinctive sentence into corporate mush and call it an improvement. Worth saying out loud: you are allowed to disagree with it. Learning to reject bad advice from a confident source is arguably the most transferable skill in this entire conversation.

What to actually say

Skip the lecture. Sit down together and try it on something real — a chapter they are struggling with, not an assignment due tomorrow. Use it as a tutor for twenty minutes. You will both learn more from that than from any policy discussion.

Then agree on a few house rules and write them down. Ours would look something like:

  • You can use AI to understand, practice, and check. Not to produce.
  • If you cannot explain it, you cannot submit it.
  • Whatever your teacher's rule is, that rule wins.
  • Never put your full name, school, address, or photos into an AI tool.
  • If something it says seems off, bring it to a person.

The part that matters more than the rules

Your kid is going to enter a workplace where using these tools well is assumed. The households that treat AI as contraband are not producing kids who avoid it — they are producing kids who use it secretly and badly.

The goal is not a kid who avoids AI. It’s a kid who can tell when it’s wrong.

That takes practice, out loud, with an adult in the room. Which is a much better use of a Sunday afternoon than another argument about screen time.