# How to Talk to Your Teen About Using AI for Homework Without a Row

**Author:** Dan Fitzpatrick
**Canonical URL:** https://parents.theaieducator.io/posts/talk-to-your-teen-about-ai-homework
**Published:** 2026-10-26T07:14:43.000Z
**Updated:** 2026-10-26T07:14:44.886Z
**Category:** Talking to Your Child

## TL;DR

Ask what your teen uses AI for, not whether they should. Sort each use into explain, check or do, and argue only about the do part. Start with one assignment, side by side, using the exact words in this post. The school's rule comes first, and no setting replaces the conversation.

## Key takeaways

- Ask "what were the last three things you asked it?" instead of "did you use AI?"
- Sort what the AI did into explain, check or do, and argue only about the do part.
- Start with one assignment, sitting beside your teen, and agree one rule for it.
- Your school's written AI rule beats any rule you set at home, so ask which teacher is clearest.
- If a row has already happened, go first: say you came at it badly and restart with one assignment.

## Article

Start by asking what they use it for, not whether they should. Most teens who use AI for schoolwork are doing something you would probably allow, and a row usually starts when a parent opens with a verdict. Pick one assignment, ask which job the AI did (explain, check or do), and agree a rule for that job only. If the school has its own rule, that comes first.

## What should I do tonight?

Choose one assignment from this week. Sit beside your teen, not across from them, and ask to be shown how it went. Not "did you use AI?" Just "how did you do this one?"

Then sort whatever you see into three jobs, using the words below, and decide together about only one of them. The rest of this page gives you the sentences to say and what to do if it goes wrong.

## Why does this turn into a row so quickly?

Because "did you use AI for this?" sounds like an accusation. It has two possible answers, and neither feels good: yes means trouble, and no means being doubted.

The numbers suggest the two sides are closer than the arguments make it seem. Pew Research Center [surveyed 1,458 US teens ages 13 to 17 and their parents between September 25 and October 9, 2025](https://www.pewresearch.org/internet/2026/02/24/how-teens-use-and-view-ai/), and published the teen findings on February 24, 2026. Of all the teens surveyed, 54 percent had used a chatbot for help with schoolwork, and about one in ten said they do all or most of their schoolwork that way. In [Pew's companion report on the parents](https://www.pewresearch.org/internet/2026/02/24/what-parents-say-about-their-teens-ai-use/), roughly six in ten parents said they would be OK with their teen using these tools to help with schoolwork. Just over half of parents said they ever talk to their teen about chatbots, and roughly four in ten said they had not had those conversations.

A second survey points the same way. Common Sense Media's [census of 1,204 US children ages 9 to 17, published June 8, 2026](https://www.commonsensemedia.org/press-releases/common-sense-media-releases-inaugural-annual-study-on-ai-use-by-tweens-and-teens), found that more than 4 in 10 children say no parent or guardian has ever talked with them about AI safety. That question was about AI safety in general, not homework, so I would not read it as the same gap. It does tell you that many children have not yet had any version of this conversation.

My reading is that most of these rows are not a disagreement about whether. Most parents are not against it, and most teens are not hiding a crime. They are two people who have never had the conversation, meeting for the first time over a laptop at nine at night.

## What can I say to start without it sounding like an interrogation?

Give up the general question. Here is the one I use instead.

**The Last Three Question: do not ask "do you use AI?" That gets a shrug. Ask "what were the last three things you asked it?" A general question gets a general answer. A specific one usually gets a phone held up.**

You could say: "I'm not checking up on you. I have no idea how people your age use this, and I would rather learn from you than guess. What were the last three things you asked it?"

Then do the hard part, which is the first sentence after they answer. Whatever the three things were, the thing you say next decides whether you hear a fourth. "Huh, I didn't know it could do that" buys you more than "you shouldn't be using it for that."

If they go quiet or bristle, you could say: "You're not in trouble. I'm trying to work out what's fair, and I can't do that while I'm guessing."

## How do we sort what the AI actually did?

Once the phone is on the table, use three words. This is my own suggested way of thinking about it, not research.

**Explain, Check, Do is a three-word way to sort what AI was asked to do with a piece of homework. Explain and check leave the thinking with your child. Do takes it away. Argue about which of the three a task was, not about whether AI is allowed.**

| Job | What it sounds like | What it leaves with your child |
|---|---|---|
| Explain | "Show me why this method works." | The thinking, with a better route in |
| Check | "Here's my answer and my working. Where did I go wrong?" | The thinking, plus feedback |
| Do | "Write the paragraph." "Solve this." | Nothing: the thinking moved |

Here is how it sorts in practice. In math, asking for a worked example of a similar problem is explain, and pasting in tonight's problem is do. In writing, asking where an argument is weak is check, and asking for the paragraph is do. In languages, asking it to quiz you on verbs is check. In science, asking why a reaction behaves as it does is explain.

It is a sorting tool and not a verdict. Explain can slide into do if the child copies the worked answer, and some teachers want certain tasks done with no help at all. For the harder cases, [the Homework Test](https://parents.theaieducator.io/posts/is-using-ai-for-homework-cheating) asks the same question from the other side: if the machine did this part, what would my child lose? If the answer is the learning, that part stays with your child.

**Argue about the job the AI did, not about the tool.** That one change takes the row out of a conversation about trust and puts it into a conversation about a task, which is something two people can actually agree on.

## What does the research say about the "do" job?

It says the worry is not made up, and also that it is more specific than "AI is bad for learning."

Hamsa Bastani and colleagues ran a field experiment with nearly a thousand high school math students and [published it in the Proceedings of the National Academy of Sciences in 2025](https://www.pnas.org/doi/10.1073/pnas.2422633122). Some students had a tutor that behaved like a standard ChatGPT interface ("GPT Base"). Others had one with prompts designed to protect learning ("GPT Tutor"). Access improved performance during practice. When the access was taken away, the students who had used the standard version did worse than students who had never had it, and the authors found the protective version largely reduced that harm.

What it does not show: it covers one subject, one kind of student and one kind of tutor, and it says nothing about essays or about your teen in particular. My interpretation is that it is a good reason to be precise about do, and not a reason to ban the other two jobs.

There is also a practical thing to try. [OpenAI's own help page](https://help.openai.com/en/articles/11780217-using-study-mode) says that ChatGPT's Study mode "can guide you with questions instead of giving the answer right away." The same page says that "there may be times when it gives a direct answer," that it can make mistakes, that it does not replace a teacher, and that students should follow their school's AI-use policies. On the web you type @study in the message box, and on the phone you tap + and choose Study. I read that page on October 26, 2026. It shows only a relative "updated" note and no fixed date, so check the wording yourself before relying on it. A setting your teen can skip is a nudge and not a lock.

## What I See in Schools

The pattern I see most often in secondary schools (that is "high school" in the US, and Years 7 to 13 in the UK) is a written policy that says AI help must be authorized, and a practice that differs from one department to the next. One teacher welcomes it for drafting. The next treats the same use as misconduct. Students notice the inconsistency long before parents do.

That is where "but everyone does it" comes from, and it is not pure invention: in [Pew's survey](https://www.pewresearch.org/internet/2026/02/24/how-teens-use-and-view-ai/), 59 percent of teens said cheating with AI happens at least somewhat often at their school.

You could say: "I know the rules are probably different in different classes. Which teacher is clearest about what's allowed?" Then ask to see the actual instructions on the assignment. If a teacher or the school has written a rule, it beats anything you agree at home.

## What if it has already turned into a row?

Restart, and go first. Parents who say "I came at that badly" tend to get a more honest answer than parents who defend the first attempt.

You could say: "I came at that badly, and I'm sorry. I'm not trying to catch you out. Can we pick one assignment and look at it together tomorrow?"

If you have found that your teen did hand in something an AI wrote, the useful next step is to help them tell the teacher before someone else raises it. That is my suggestion, and the school's process comes first. If the situation runs the other way and the school has accused your child, [that is a different situation with its own steps](https://parents.theaieducator.io/posts/child-accused-of-using-ai).

If your teen seems overwhelmed rather than lazy, the real conversation may be about workload or worry, and their teacher, the school counselor or their doctor is the right next call.

## How does this change by age?

**Ages 9 to 12** (US middle school and upper elementary; Years 5 to 8 in the UK). Most chatbots set 13 as their own minimum age, so [check the age rules for the tool in question](https://parents.theaieducator.io/posts/what-age-can-kids-use-ai). If a child this age uses AI for homework, keep it in the room with you or inside a school-provided tool. Stay with explain. Use the checking questions from [my guide to explaining AI to a child](https://parents.theaieducator.io/posts/how-to-explain-ai-to-a-child) in this order: Who says? How could we check? What would it look like if this were wrong?

**Ages 13 to 15.** These are the row years. Use the Last Three Question, then Explain, Check, Do, one assignment at a time. Keep it easy to talk to you. If this also touches how they use AI for company or comfort, [the Easier Door](https://parents.theaieducator.io/posts/my-teen-talks-to-ai-about-feelings) is the question I would ask there: is talking to me easier than talking to the app?

**Ages 16 to 18.** Autonomy matters more than supervision. Ask them to explain their own rule to you. You could say: "Tell me your rule for using it, and what you would say if a teacher asked how you work." This is also the age where exam and coursework rules, and later application essays, start to count. [Can colleges tell if a teen used AI on an essay](https://parents.theaieducator.io/posts/can-colleges-tell-if-my-teen-used-ai-on-their-essay) covers that.

## What none of this can promise

No script guarantees an honest answer, and no setting will do the talking for you. Pew's figures are what teens and parents reported about themselves, not what they were observed doing. The Bastani study covers math. Rules at your child's school come first, and so does any professional advice you have been given about your child.

What the conversation can do is make it likelier you will hear about it when something goes wrong. That is worth more than a rule you cannot check.

If your child's school has not spoken to parents about AI yet, this is the work I do with schools, and it is worth sending to your head of year or principal: [my AI training for schools](https://theaieducator.io/ai-training-for-schools?utm_source=parents.theaieducator.io&utm_medium=referral&utm_campaign=talk-to-your-teen-about-ai-homework). If you would like to keep up without the panic, I write [a newsletter](https://theaieducator.io/?utm_source=parents.theaieducator.io&utm_medium=referral&utm_campaign=talk-to-your-teen-about-ai-homework#newsletter) that parents are very welcome to join.

*Dan Fitzpatrick is a former teacher, assistant headteacher and school trustee who now trains educators on AI. [About Dan](https://theaieducator.io/about?utm_source=parents.theaieducator.io&utm_medium=referral&utm_campaign=talk-to-your-teen-about-ai-homework).*

## Frequently Asked Questions

### Should I read my teen's chat history?

Usually you cannot, and I would ask them to show you instead. The controls on the major chatbots mostly do not show a parent the conversations themselves, so asking "what were the last three things you asked it?" gets you further than hunting. See the earlier post on what parents can and cannot see for the detail on ChatGPT.

### What do I say when my teen says everyone uses AI for homework?

Agree that it is common, then bring it back to their own work. In Pew Research Center's survey of US teens, published February 24, 2026, 59 percent said cheating with AI happens at least somewhat often at their school. You could say: "I believe you. I'm asking what is fair for you on this assignment."

### Is it fine for my teen to use AI to check their work?

Yes, if their school allows it, because checking leaves the thinking with them. The test is whether they did the work first and then asked where it went wrong. If the AI is producing the first draft, that is the do job, and it needs the teacher's say.

### Should I just ban AI for homework?

A ban you cannot see is a rule you cannot check, so I would start with the conversation and add a ban only if your school or your own judgment calls for one. If the school has banned it for a task, that rule comes first, and your teen can be told so plainly.

### What if my child's school allows AI for homework?

Then the school's rule decides what is permitted, and your conversation moves to what your child is learning. Ask the teacher which tasks they want done without help. The Explain, Check, Do sort still works for deciding which part is yours to talk about at home.

### How is this different for a child under 13, or in the UK?

Most chatbots set 13 as their own minimum age, so a child of 9 to 12 should use AI only alongside you or inside a school-provided tool. In the UK the same conversation applies to secondary school students, usually from Year 7, and the school's own rules or the exam board's coursework rules come first.