Should Your Child Use ChatGPT for 11+ Prep? What Actually Helps
Here's a number worth sitting with: 64% of UK children are already using AI chatbots, and 42% of 9-17 year-olds who've tried one have used it specifically for schoolwork. This isn't a "should we allow it" conversation anymore — it's already happening in most households, quietly, on a device your child probably already has access to. The useful question isn't whether to ban it. It's where it genuinely helps 11+ preparation, and where it does something worse than nothing.
That distinction matters more for the 11+ than almost any other exam a child sits, for one specific reason: the 11+ is designed to test a child's own reasoning, worked out under time pressure, not their ability to retrieve a correct answer. An AI tool that gives the answer away doesn't just fail to help — it actively removes the exact practice the exam is testing.
Quick summary
Quick tips:
- Use AI to get a concept explained differently — never to get a practice answer.
- Treat every AI answer as unverified until checked, especially for VR/NVR and Maths.
- If a child never got the wrong answer, they never actually practised.
The numbers, laid out plainly
Where AI genuinely helps 11+ preparation
Explaining a concept a different way. If a child doesn't understand why a fraction question works the way it does, asking an AI tool to explain it more simply — or with a different example — is a genuinely useful use of the technology, similar to asking a second tutor for a fresh explanation.
Simplifying dense text. For a child who finds a comprehension passage's vocabulary overwhelming, having a chatbot rephrase a difficult paragraph in simpler language can help them access the underlying idea before returning to the real, unsimplified passage — useful as a stepping stone, not a substitute.
Generating extra reading material or vocabulary lists. Asking for age-appropriate passages on a topic a child is interested in, or a list of synonyms to learn, is a reasonable, low-risk use that doesn't touch the exam's actual skill.
Where it quietly undermines 11+ prep specifically
Solving practice questions for them. This is the core issue. The 11+ doesn't reward knowing an answer — it rewards being able to work one out, under time pressure, without help. A child who gets AI-generated answers to practice questions is optimising for the wrong thing entirely: a stack of correct answers with none of the underlying skill that produces them on exam day.
Verbal and Non-Verbal Reasoning specifically. VR and NVR test fast, systematic pattern-spotting — a skill built only through a child's own repeated attempts, including the wrong ones. An AI chatbot can explain the rule behind a puzzle after it's shown the answer, but it cannot build the instinct of spotting that rule independently under a ticking clock, which is the actual thing being tested.
Confident wrong answers. Research into children's AI chatbot use has found many never question the accuracy of what they're told — a chatbot delivers an incorrect answer with exactly the same confident tone as a correct one. For a mark-scheme-based exam where there's one right answer, this is a genuine risk, not a hypothetical one.
Skipping the productive struggle. Getting a question wrong, working out why, and trying a similar one again is the actual mechanism through which 11+ skills improve. An AI shortcut that removes the "getting it wrong" step removes the learning, even if the child ends up with more correct-looking homework.
A simple test for any AI use
Before a child uses an AI tool as part of 11+ prep, one question separates the useful cases from the harmful ones: is this explaining something, or is this doing the exam-relevant thinking for them? "Explain why 3/4 is bigger than 2/3" passes that test. "What's the answer to this practice question" doesn't — even if the child intends to "just check," because checking before genuinely attempting removes the productive struggle the exam is built around.
Parent advice
Sit with the AI use the first few times, not just the first time. A single supervised session doesn't tell you what happens on session ten, once the novelty (and the temptation to shortcut a hard question) has settled in.
Ask your child to explain the answer back to you, not just show it. If they got help from an AI tool and genuinely understood the explanation, they should be able to walk you through it in their own words. If they can't, the "help" didn't actually help.
Be aware of the "friend" dynamic. With over a third of children reporting that chatbot conversations feel like talking to a friend, it's worth an occasional check-in on how your child is using AI tools generally, not only for schoolwork.
Student tips
If you're stuck on a practice question, try it properly first — even a wrong guess tells you something useful about where the gap is. If you're still stuck after a real attempt, ask an AI tool (or a parent, or a tutor) to explain the idea behind the question, not the answer to it. Then try a similar question yourself, without help, straight after — that's the step that actually makes it stick for exam day.
Tutor notes
- Ask directly about AI use in your first conversation with a family. Many parents haven't thought about it as a distinct category from "using the internet" — naming it explicitly surfaces habits worth addressing early.
- Watch for suspiciously fast, suspiciously polished independent work. A sudden jump in written work quality without a matching jump in verbal explanation ability is a common tell.
- Model good AI use rather than banning it outright. Demonstrating "ask it to explain, then try the question yourself" as a habit is more durable than a blanket rule a child works around unsupervised.
So what should you actually do?
Set one household rule and stick to it: AI can explain, never answer. If your child brings a hard question to an AI tool, the request should always be framed as "help me understand this," never "what's the answer" — and any explanation given should be followed immediately by your child attempting a similar question alone, unaided, to prove the understanding actually transferred.
The deeper reason this matters: exam-day confidence comes from a child having genuinely done the work themselves, over and over, including the questions they got wrong along the way. A stack of AI-assisted "correct" practice answers can look like progress while quietly leaving a child less ready than before — because the one thing that builds real exam-day calm, repeated independent practice, never actually happened.
Frequently asked questions
How many children actually use AI chatbots like ChatGPT?
A UK survey of 1,000 children and 2,000 parents found 64% of children are using AI chatbots for tasks ranging from homework help to emotional support and companionship, and 42% of children aged 9-17 who have used an AI chatbot have specifically used one to support schoolwork.
Is it OK for my child to use ChatGPT to help with 11+ practice?
It depends entirely on how it's used. Asking an AI chatbot to explain a concept in simpler words, or to generate extra practice on a topic, can genuinely help. Asking it to solve a practice question and give the answer defeats the purpose of 11+ preparation specifically, because the exam tests a child's own reasoning under time pressure, not their ability to look up an answer.
Can AI chatbots give wrong or misleading answers?
Yes. Research has highlighted that children frequently never question the accuracy of what a chatbot tells them, and chatbots can present incorrect or oversimplified information with the same confident tone as correct information — which is a particular risk for exact, mark-scheme-based subjects like Verbal and Non-Verbal Reasoning.
Why can't AI replace practising Verbal and Non-Verbal Reasoning?
VR and NVR questions test a child's own pattern-spotting speed and systematic reasoning under exam time pressure — skills built through repeated, active practice. An AI chatbot can explain the rule behind a puzzle after the fact, but it can't build the fast, independent pattern-recognition instinct the real exam actually measures; only the child's own repeated practice can do that.
Do children see AI chatbots as more than just a tool?
Concerningly, yes for a meaningful minority. Over a third (35%) of children who use AI chatbots say talking to one feels like talking to a friend, and six in ten parents say they worry their child believes the chatbot is a real person — a dynamic worth being aware of if a child is using one unsupervised for revision help.
Related SharpY Mind resources
Where to go from here
SharpY Mind explains every mistake with a worked solution the moment it happens — the "why," not just the answer — so your child builds real understanding through their own practice, not a shortcut around it.
⚡ Try SharpY Mind free