Can my child under 13 use Google Gemini?

Yes, if you allow it. Google’s own For Families page, read 31 October 2026, states you can give a child under 13 access to Gemini Apps through a supervised Google Account, managed in Family Link under Controls, then Gemini. Separately, a school-issued education account is governed by the school rather than your Family Link settings — the route most parents do not realise is open.

Can I see what my child asks Gemini?

Family Link governs access, not conversations. The practical controls are the switch and the rules you set about where and when it is used. If seeing what your child actually does with AI matters to you, that is a reason to choose a product built with a parent side rather than to expect one from a general-purpose assistant.

Is Gemini safe for a 9-year-old?

Safer with you in the room, and Google does not claim otherwise. Its own page says the filters "aren't perfect", that Gemini can present inaccurate information as factual, and that a child should be told it is not a person to confide in. Read that as a general-purpose tool children are permitted to use rather than one designed for them — reasonable to allow with supervision, poor to leave running unattended.

Is Gemini or ChatGPT better for a child?

The wrong comparison to spend effort on. Both are general-purpose assistants for adults with child access bolted on, both warn their filters are imperfect, and neither shows you what your child learned. The question that separates products usefully is whether anything exists afterwards — something made, something explained, something you can see — and on that axis the two are much more alike than different.

Parent’s guide

Is Google Gemini safe for kids?

Gemini arrives differently from the other chatbots: not because your child signed up for it, but because it is already inside the Google account the school gave them, or one toggle away in Family Link. The most useful source on whether it is safe is not a review site — it is Google’s own page for parents, which says three things more candid than most of the coverage. Here is what it says, what it leaves out, and the four decisions that are actually yours.

Flat pastel illustration of a blue speech bubble pouring into a green funnel with a slotted mesh across its top, three coloured shapes caught inside, one dark shape falling through the spout below it, and an orange shield with a white tick standing alongside, on a cream background.

What Google itself tells parents

Read on 30 October 2026, Google’s own For Families help page says a parent can give a child under 13 access to Gemini Apps through a supervised account, controlled in Family Link. So the first thing to know is that this arrives as a switch — and that the switch is most of what the switch does.

Then it says three things worth quoting, because they are more honest than the marketing around them. On filtering: "Our filters try to limit inappropriate content from appearing, but they're not perfect. Your child may encounter content you don't want them to see." On accuracy: Gemini "can hallucinate and present inaccurate information as factual", and parents should "help your child think critically about Gemini Apps responses". And on what it is: a child should understand "it's an AI tool, not a person to confide in", and that it "should not be treated as a friend or used to discuss personal problems or secrets."

Take that at face value and the headline question is most of the way answered. Google is not claiming the filter is reliable, not claiming the answers are accurate, and is specifically warning against the use case — confiding — that worries parents most. That is a vendor describing a general-purpose tool children are permitted to use, not a product built for children. It is also a vendor page, which is its own limit: it tells you what Google says on the day you read it, not what the product does, and it can change with no notice. Treat the quotes above as dated.

The four decisions that are actually yours

Most of what gets written about this is a verdict on the product. The useful frame is narrower: here is what you can actually control, and what each control does not reach. The right-hand column is the part the reviews tend to skip.

The decision	How you make it	What it does not cover
Whether they have access at all.	Family Link: Controls, then Gemini. One switch, on or off, for a supervised account.	A school-issued Google Workspace for Education account is governed by the school, not by your Family Link. Ask the school which AI is enabled on the device they sent home.
Whether they use it alone.	Your rule, not a setting. Google’s own advice is to use it together at first.	No control enforces this, and none tells you afterwards whether it happened. Sitting alongside for the first few sessions is doing the work that no toggle does.
Whether they treat it as a person.	A conversation, repeated. Google’s page says explicitly it is "not a person to confide in".	Nothing in the product prevents a child using it that way, and the warmth of the replies works against you here. This is the risk with the weakest technical mitigation.
Whether they believe what it says.	The one habit worth drilling: ask how it knows, and check one thing it claimed.	There is no setting for accuracy. Google says it can present inaccurate information as factual, which makes verification a skill your child needs rather than a feature you can switch on.

What the independent picture adds, and what it cannot settle

Common Sense Media’s Youth AI Safety Institute publishes independent risk assessments of the major chatbots, and the reporting around their Gemini assessment in 2025 is where the "high risk" framing in most articles comes from. The mechanism in their work that matters most for a general-purpose assistant is the transfer of trust: a tool that is genuinely good at homework help earns a credibility it then spends on questions it handles badly.

Two honest caveats on using that. Their assessments are written primarily about teen experiences rather than about seven-to-twelve-year-olds, and a risk rating is a judgement rather than a measurement. So it belongs in your thinking as a reason to supervise, not as a number to quote at a school.

There is also a question none of this answers, and it is the one COPPA exists for: what happens to what your child types. The amended COPPA Rule sets a floor for US operators — verifiable parental consent, limits on retention, a broader definition of personal information — but it is a legal floor rather than a quality bar, and a product can sit comfortably above it while still being a poor place for a nine-year-old to spend an hour. The question worth asking of any AI your child uses is not "is it legal" but "what does it do with what my child tells it, and can I see any of it?"

The same question for ChatGPT, with the risks set out by age · AI apps that have a parent side at all

If you are turning it on, do these five things first

1. Find out what the school already enabled Before you decide anything at home, ask which AI tools are on for the account on the school device. Plenty of parents are carefully weighing a Family Link switch for an account their child does not use, while a school Workspace account with different settings sits in the other tab.
2. Use it together for the first week Google's own recommendation, and the only one on this list that reliably changes what happens. You are not supervising so much as showing: what a good question looks like, what you do when the answer seems off, that it is normal to disagree with it.
3. Say the "not a person" sentence out loud, early Google puts this in writing for a reason. The sentence that works at this age is concrete: it does not remember you, it does not know you, and it will agree with you because that is how it was built — so it is a bad place for anything you would only tell a friend.
4. Drill one habit: how do you know? Pick one claim in any answer and check it together, once a week. It is the only defence against hallucination that survives you leaving the room, and it is a skill that transfers to every other source they will ever read.
5. Decide what it is for in your house The difference between a tool and a habit is whether there is an answer to "what do we use this for". Explaining something a teacher said, planning a project, checking spelling — fine, and specific. "Whenever you want" is how a general-purpose assistant becomes the default place a child goes with everything, including the things it is worst at.

The sources under this page

One vendor page quoted directly and dated, one independent assessment body, and the US legal floor. Note that the strongest source here — Google's own words — is also the one most likely to change without notice, which is why the date is printed next to it.

Google For Families Help, read 31 October 2026 — Guide your child’s Gemini Apps experience. (Google’s own page for parents: it states that a parent may give a child under 13 access via a supervised account, that "Our filters try to limit inappropriate content from appearing, but they’re not perfect", that Gemini "can hallucinate and present inaccurate information as factual", and that a child should understand "it’s an AI tool, not a person to confide in". Its reporting section adds that if a child "encounters inappropriate or incorrect content, you or your child can report it", and that "You can report it by submitting feedback or your child can mark it as a bad response".) The direct quotes in the first section, re-read 31 October 2026 (all four still live, exact-count verified): under-13 access via a supervised account, filters that are "not perfect", hallucination stated plainly, and the "not a person to confide in" advice. A vendor page, so it is evidence about what Google tells parents rather than about what the product does. What it does not show: A vendor’s own help page, not an independent assessment, and it describes the product on the date it was read. It is strong evidence about what Google itself tells parents and weak evidence about what the product actually does. On the reporting half specifically: the page names a route and makes no claim at all about what happens after a report is filed, so nothing here supports telling a parent that reporting will change the model’s behaviour — only that the vendor asks to be told.
Common Sense Media, Youth AI Safety Institute risk assessments — Independent risk assessments of major AI chatbots for young people, including the finding that competence at homework help leads teens and parents to over-trust the same systems on things they handle badly. Used for the independent half and for the transfer-of-trust mechanism — competence at homework help buying credibility on questions handled badly. Written mainly about teen experiences, so it is a reason to supervise rather than a rating to quote about a nine-year-old. What it does not show: Assessments are written for teens, and the general-purpose chatbots tested are not the kid-specific products this page is about. The transfer-of-trust mechanism is what carries over to ages 7–12; the ratings do not.
FTC, amended COPPA Rule — The Children’s Online Privacy Protection Rule as amended (published 22 April 2025; full compliance required from 22 April 2026), adding separate verifiable parental consent for third-party disclosure, data-retention limits, and a broader definition of personal information. Used only to make the point that the legal floor is a floor. Compliance tells you something about data handling and nothing about whether this is a good way for a child to spend an hour. What it does not show: A legal floor for US operators, not a quality bar. A product can be fully COPPA-compliant and still be a bad way for a nine-year-old to do maths.

Questions parents ask

Can my child under 13 use Google Gemini?: Yes, if you allow it. Google’s own For Families page, read 31 October 2026, states you can give a child under 13 access to Gemini Apps through a supervised Google Account, managed in Family Link under Controls, then Gemini. Separately, a school-issued education account is governed by the school rather than your Family Link settings — the route most parents do not realise is open.
Can I see what my child asks Gemini?: Family Link governs access, not conversations. The practical controls are the switch and the rules you set about where and when it is used. If seeing what your child actually does with AI matters to you, that is a reason to choose a product built with a parent side rather than to expect one from a general-purpose assistant.
Is Gemini safe for a 9-year-old?: Safer with you in the room, and Google does not claim otherwise. Its own page says the filters "aren't perfect", that Gemini can present inaccurate information as factual, and that a child should be told it is not a person to confide in. Read that as a general-purpose tool children are permitted to use rather than one designed for them — reasonable to allow with supervision, poor to leave running unattended.
Is Gemini or ChatGPT better for a child?: The wrong comparison to spend effort on. Both are general-purpose assistants for adults with child access bolted on, both warn their filters are imperfect, and neither shows you what your child learned. The question that separates products usefully is whether anything exists afterwards — something made, something explained, something you can see — and on that axis the two are much more alike than different.

A different shape of AI for a nine-year-old

A general assistant answers whatever gets typed into it. Luchi is built the other way round: kids make things, explain what they made, and you get a short weekly note in their own words — so "what did they do with it" has an answer without anyone reading a transcript.

Start your first 7 days for $5

Is AI safe for kids?

Part of the series: Is AI safe for kids?

Related guides in this series

Is ChatGPT safe for kids? — A straight answer by age — the risks that actually come up, and the five habits that make AI safe when you do let kids use it.
Screen time and AI — Not all screen time is equal. Why active, creative AI use is different from passive scrolling — and how to keep your child on the right side of the line.
What is my kid doing on AI? — You can’t read every transcript. Four checks that take under a minute each, what each one tells you, and the one question a ten-year-old actually answers.
AI apps a parent can actually monitor — Alert, transcript, or record — three different things sold under one word. What seven products say they show a parent, quoted from their own sites.
The school Chromebook already has AI on it — A school account is managed by the district, not by your Family Link. Who actually holds the switch, what Google’s own documentation says a student account gets, and five questions worth one email.
Deepfakes: what to actually teach a 10-year-old — The “look closely, count the fingers” advice has been measured, and it does not work — neither warnings nor money improved anyone’s accuracy. What to teach instead, and why a detector app is not the backstop.
What to ask at a parent-teacher conference about AI — The list you found ends with “ask how you can help with homework” — the one form of involvement a 50-study meta-analysis found was not associated with achievement. Six questions that are, and what each answer tells you.
Is Siri the kind of AI parents worry about? — Apple has quietly split Siri in two — its own page now says Apple Intelligence “powers an entirely new version of Siri: Siri AI”. How to tell which one your child is talking to, what Apple keeps, and the settings worth changing tonight.
A family AI agreement you will actually keep — Twenty clauses, signed on Sunday, dead by Thursday. A four-clause version built on what a 94-test meta-analysis found actually makes a plan start — naming the when and the where — plus the one question to ask each week.
An AI said something it should not have — Google’s own parent page says the filters "are not perfect". The order to do things in — keep the conversation, ask the one question, report from inside the chat, decide one thing — and the point at which this hands you to somebody qualified.
My child believes everything the AI says — The gap is usually confidence rather than method, and a check is still a looking-up — which across nine experiments left people MORE sure of what they understood. The calibration habit that moves it, in four turns.