Revision

Can My Child Use ChatGPT for Homework? What UK Schools Actually Allow

Your child says everyone uses it and the teachers don't mind. Some of that may be true. Here is what the rules in England actually say, where the real line sits, and a two-minute check that tells you whether AI is helping your child learn or doing the learning for them.

TutorMint FounderPublished 29 September 202611 min read
A parent holding a mug listens as a teenager in school uniform explains their homework from an open exercise book, with the laptop closed on the kitchen table

The short version

  • There is no national rule on AI for homework in England. The Department for Education leaves it to each school, so the answer is in your child's school policy, which many parents have never been shown.
  • ChatGPT's own terms say users must be at least 13, and anyone under 18 needs a parent's permission. For Year 7 and most of Year 8 the answer is not yet. After that, it is your decision first.
  • Homework and coursework are different. Ordinary homework follows school rules. Coursework that counts towards a GCSE or A-Level comes under exam-board rules, and AI misuse there is malpractice that can lead to disqualification.
  • In the National Literacy Trust's 2026 survey, 1 in 4 teenagers said that when they used AI for homework, they usually just copied what it told them. That is the use to stop. Asking it to explain a step, set practice questions or comment on their own draft is a different thing.
  • One two-minute check tells you which kind your child is doing: close the laptop and ask them to explain their answer back to you.

The homework comes back with a comment you didn't expect. Or it was finished in eleven minutes and reads better than anything your child has written before. Or you ask how the history essay is going and are told that everyone uses ChatGPT, that the teachers know, and that it's fine.

Some of that may be true. The difficulty for parents is that nobody has told them which parts.

This guide sets out what the rules actually are in England, where the line that matters sits, and how to tell in two minutes whether AI is helping your child learn or doing the learning for them.

Is it allowed? It depends on the school

There is no national rule on using AI for homework. The Department for Education's policy paper on generative AI in education says schools are free to make their own choices about how AI is used, as long as they meet their wider duties such as safeguarding. Some schools only let teachers use it. Others allow pupils to use it, but only in certain subjects or year groups.

“Schools and colleges may wish to review homework policies, and other types of unsupervised study to account for the availability of generative AI. This may include developing guidance on when it is acceptable or appropriate to use generative AI tools for educators, students and pupils.”

Department for Education, Generative artificial intelligence (AI) in education, updated 12 August 2025

Note the word "may". Reviewing the homework policy is a suggestion, not a requirement, and so is the line that follows it, that schools may wish to consider how they talk to parents about AI. In practice, one school bans AI for homework outright, the next encourages it for revision, and a third has a policy on its website that nobody has mentioned to you. Two children in the same street can be working under opposite rules.

Teachers are not settled on it either. In the National Literacy Trust's 2026 survey, 60.1% of teachers said they were worried about pupils using AI, up from 43.2% a year earlier. Among secondary teachers the figure was 67.7%.

The quickest way to find out: search your child's school website for "AI policy" or "artificial intelligence", and read the homework policy. If you find nothing, email the head of year and ask. It is a reasonable question, and a clear answer protects your child as much as it informs you.

What ChatGPT's own age rules say

Before the school's rules come the app's. OpenAI's terms for UK users say you must be at least 13 to use ChatGPT, and that anyone under 18 needs a parent or guardian's permission and should ask them to read the terms together. Other chatbots set their own age limits, and the Department for Education tells schools to comply with them.

So the first decision is yours, not the school's. Here is how it maps onto school years:

ChatGPT's terms by school year in England

Year 7

Typical age
11 to 12
ChatGPT's own terms
Below the minimum age of 13
Who decides about homework
No account, whatever the school allows

Year 8

Typical age
12 to 13
ChatGPT's own terms
Only from their 13th birthday, with your permission
Who decides about homework
You first, then the school's policy

Years 9 to 11

Typical age
13 to 16
ChatGPT's own terms
Allowed with your permission
Who decides about homework
The school's policy, plus exam-board rules for coursework

Years 12 and 13

Typical age
16 to 18
ChatGPT's own terms
Allowed with your permission
Who decides about homework
The school or college's policy, plus exam-board rules for coursework

Terms are not a lock. Nothing stops a 12-year-old typing in a false birthday, and AI assistants now sit inside apps your child already uses. But if a Year 7 child is using ChatGPT for homework, the service itself says they shouldn't be, which is a useful fact to have when the conversation starts.

How teenagers are actually using it

It's easy to picture every teenager handing their homework to a chatbot. The best UK data is more mixed than that. The National Literacy Trust surveyed 40,543 young people aged 13 to 18 in early 2026, and 79.0% said they had used AI.

1 in 4

teenagers say that when they use AI for homework, they usually just copy what it tells them

24.1% in 2026, against 25.1% in 2025, rising to 31.6% among young people on free school meals. National Literacy Trust, Young people's use of generative AI to support literacy in 2026.

That is the number to worry about, and it has barely moved in a year. The rest of the picture is more encouraging. More than half (56.5%) said they usually add their own thoughts to what AI gives them, and 52.3% said they check its answers because they might be wrong, both clearly up on 2025. And 62.8% agreed that you should use AI to help you think, not to think for you.

One young person in the charity's focus groups put it better than most policies. They had stopped letting AI do their work, because "I can't bring it to the exam anyway."

Parents are divided too. In a separate survey of 3,073 UK parents for the same charity, 35.9% agreed that AI is not something they want their child to use, while 32.9% disagreed. Among parents of 12 to 15-year-olds, 33.7% said they had used AI with their child to help with homework. If you are unsure where you stand, you are in good company.

Homework and coursework are not the same thing

This is the distinction that matters most, and the one most children have not been told clearly.

Ordinary homework, such as a worksheet, a practice essay or a set of maths questions, follows your school's rules. If a child copies from AI, the usual result is a conversation with the teacher and a lesson they didn't learn.

Work that counts towards a GCSE or A-Level grade is different. Most marks come from exams sat under supervision with no internet, so AI is not available there at all. But some subjects include coursework, which exam boards call non-exam assessment (NEA). GCSE design and technology is half coursework, and several A-Levels, including history, English literature and geography, have an NEA component. The Extended Project Qualification is coursework from start to finish. All of this is governed by the Joint Council for Qualifications (JCQ), which sets the rules for the main exam boards.

JCQ's guidance on AI in assessments is blunt. Among other things, it counts as misuse:

  • Copying or paraphrasing sections of AI-generated content, so that the work is no longer the student's own
  • Using AI to complete parts of the assessment, so that it doesn't reflect the student's own work, analysis, evaluation or calculations
  • Not acknowledging AI when it has been used as a source of information, or acknowledging it poorly

Misuse is malpractice. The sanctions available include disqualification and being barred from taking qualifications for a number of years. And even where a student openly acknowledges the AI, JCQ says the copied sections will not be rewarded, because they don't show the student meeting the marking criteria themselves. Put simply, acknowledged AI text scores nothing, and unacknowledged AI text can cost the whole qualification.

These are not hypothetical. JCQ's guidance includes real, anonymised cases. In one, two A-Level History students used ChatGPT on their coursework. One was disqualified from the A-Level, and the other lost every mark for the coursework component. Both had signed a declaration that the work was their own.

If your child is in Year 10 or above and has any coursework on the go, ask the teacher exactly what is and isn't allowed for that piece. The rules for that work are far stricter than for anything else they do at home.

Uses that help, and uses that replace learning

Take the rules away and a simpler question is left: is the AI doing the thinking, or helping your child do it? The same tool can do either. It depends on when it is used and what it is asked to do.

Usually helps learning

  • Asking for a different explanation of a step they have already tried and got stuck on
  • Asking for practice questions on this week's topic, then answering them without help
  • Getting feedback on a draft they wrote themselves, then doing the rewrite themselves
  • Asking it to test them on flashcards or key terms
  • Checking the meaning of an unfamiliar word in something they're reading

Usually replaces learning

  • Pasting in the homework question and copying the answer
  • Asking it to "improve" an essay and handing in the new version
  • Reading an AI summary of the class novel instead of the book
  • Getting every step of a maths problem before trying it
  • Using it on anything that counts towards a grade without checking the rules

Summaries deserve a special mention. In the 2026 survey, 53.5% of teenagers who used AI at least monthly said they had used it to shorten or summarise something to make it quicker to read, up from 33.0% a year before. For a class novel, the reading is the homework. Our post on children who hate reading explains why that reading matters so much in Key Stage 3.

There is also a plainer problem: chatbots get things wrong. OpenAI's own terms warn that output may not always be accurate and should not be relied on as a sole source of truth. A general chatbot doesn't know which exam board your child sits, which method their teacher wants to see, or how a mark scheme gives out marks. Only 16.4% of teenagers in the 2026 survey agreed that you can trust the information you get from AI, which is a sensible instinct.

How people learn points the same way. Pulling something out of your own memory is what makes it stick, which is why the techniques in our guide to active recall and spaced repetition work, and why reading an answer, however good, does so little.

The two-minute check: can they explain it back?

You don't need to understand the technology, or even the subject, to tell which kind of use your child is making. Pick one piece of finished homework and ask, casually, for a quick walk-through with the screen closed.

Probably using it well

  • Explains the answer in their own words
  • Can tell you where they got stuck and what helped
  • Can do a similar question in front of you
  • Their writing still sounds like them

Worth a closer look

  • Can't explain an answer they wrote an hour ago
  • Uses words or methods they can't define
  • Can't start a similar question without the screen
  • Their writing suddenly sounds like a textbook, perhaps with American spellings

If most of what you see is in the right-hand column, don't start with a ban. Start by asking why. A child who copies every night is usually a child who is stuck, and AI has become the quickest way to make the stuck feeling go away. The copying is the symptom.

Five questions to ask the school

These fit in one email to the form tutor or head of year:

  1. 1Does the school have a written policy on pupils using AI, and where can I read it?
  2. 2Is AI allowed for homework, and does that differ between subjects?
  3. 3Which of my child's subjects include coursework that counts towards the final grade?
  4. 4What are the rules on AI for that coursework, and how should my child acknowledge any use?
  5. 5Is there an approved AI tool the school would rather pupils used?

Ground rules that work at home

Every family will land somewhere different. These suit most homes with a child aged 13 or over:

  • Try it first. One honest attempt before any AI, so there is something real to ask about.
  • Ask for help, not answers. "Explain this step another way" is fine. "Do question 4" is not.
  • Use it in a shared room. Not as surveillance. It's the same reason you might not want homework done in bed.
  • Say when you used it. If AI helped, the teacher should know. It's a good habit now, and a requirement for coursework later.
  • Coursework is different. Nothing that counts towards a grade goes near AI until the teacher's rules have been checked.

If homework is already a nightly fight, our guide to homework battles covers how to help without taking over, and most of it applies here too.

Where a tutor fits

A good tutor won't do your child's homework either. What a tutor can do, and a chatbot can't, is notice why your child is stuck. A chatbot answers the question it is asked. A tutor sees that the same kind of question has gone wrong three weeks running, works out that the real gap is two topics back, and teaches that until your child can do it with the laptop closed.

That matters most for the child who copies because they are lost. AI can hide a gap like that for a long time, right up until the mock exam, where there is no chat window.

Our online homework help is one-to-one, built around the work your child has been set. We teach the method and never hand over the answers, so your child finishes the session able to do the next one alone. Lessons are one hour and start from £15 at KS3, £20 at GCSE and £25 at A-Level, with no contract and a free demo class first. See KS3, GCSE and A-Level tutoring for how lessons work at each stage.

Ready to find out where your child stands?

TF

Written by TutorMint Founder

10+ years of classroom and one-to-one teaching experience. Facts checked 29 September 2026.

Found this useful?

Add Tutormint as a preferred source and our guides show up higher when you search Google.

Common Questions

Frequently Asked Questions

It depends on how it is used and what the school allows. There is no national rule in England; each school sets its own. Copying an AI answer and handing it in as your own work is treated as cheating almost everywhere. Using AI to explain a step, set practice questions or give feedback on your own draft is allowed by many schools, but check the school's policy and tell the teacher when AI was used.

OpenAI's terms for UK users say users must be at least 13, and anyone under 18 needs a parent or guardian's permission. In school terms, that rules out Year 7 and most of Year 8. Other AI chatbots set their own age limits, so check the terms of any app your child uses.

Often, yes. Teachers know how your child writes, and a sudden change in style, vocabulary or spelling is a common warning sign. For coursework, schools and exam boards also use AI detection software, and the malpractice cases JCQ has published include students identified this way. JCQ also notes that detection tools vary in accuracy, which is one reason teachers usually follow up by asking students to explain their work.

If AI-generated work is submitted as the student's own, it counts as malpractice under JCQ rules, which cover the main exam boards. Sanctions range from a warning to losing marks, disqualification from the qualification, or being barred from exams for a period. Even acknowledged AI content earns no marks, because it doesn't show the student's own work.

A ban is hard to enforce, since AI is built into many apps children already use, and it doesn't teach them to use it well. For most families with a child aged 13 or over, clear rules work better: an honest attempt first, AI for explanations rather than answers, and nothing near coursework without the teacher's rules. For a child under 13, ChatGPT's own terms already say no.

Yes, if the reliance comes from being stuck. A tutor finds the gap behind the copying and teaches it until your child can do the work alone. Tutormint's lessons are one-to-one and online, one hour long, from £15 at KS3, £20 at GCSE and £25 at A-Level, with a free demo class first.

Start With a Free Demo Lesson

No card details, no obligation — just an honest read of where your child stands.

WhatsApp Us a Question
WhatsApp Us