Last updated November 7, 2026. Microsoft's pages checked that day.
Microsoft Copilot is a reasonable tool for a teenager aged 13 or older, with an adult close by, and it is not a chatbot for a younger child. The catch is that there are two of them. The Copilot on your family's Microsoft account follows rules you can adjust; the Copilot on a school account follows rules the school sets. Find out which one your child is using before anything else.
Which Copilot is my child actually using?
This is the question that decides the rest, and most families have never been asked it. The same name sits on two quite different products, and the account a child signs in with picks between them.
Do not ask whether Copilot is safe. Ask which Copilot, signed in with which account, and who set the rules.
Here is the comparison, checked against Microsoft's own pages on November 7, 2026.
| Signed in with | Who sets the rules | Minimum age | What a parent can do |
|---|---|---|---|
| A personal Microsoft account (the Copilot app or website) | Microsoft's young-people rules, plus a parent through Microsoft Family Safety | 13, and higher in some countries | Block the app or the website, set time limits, choose features such as image generation, voice, group chat and AI companions |
| A school account, in Microsoft 365 Copilot Chat | The school's IT administrator | Teen protections apply to ages 13 to 17 where the school has recorded the student as a teen | Nothing directly. The school decides who gets access |
| A child account below the statutory age, in Family Safety | Microsoft's AI controls for families | AI features default to off, and the parent cannot switch them on | Nothing to switch on. The default protects |
The sources are Microsoft's own: its Copilot for young people page for the current app, its Family Safety AI controls page, and a Microsoft 365 message center notice (MC1509805) for school accounts. I could read that notice only through a public mirror, not in a school admin center, so treat the school row as the best public reading and not the last word.
What does Microsoft say is different for under-18s on a personal account?
The current Copilot page says you must be at least 13 and that "in some countries or regions, the minimum age may be higher depending on regional requirements." For young people, it says Copilot "doesn't personalize results based on your conversation history" and does not show personalized ads, though you may still see ads based on what you ask in that chat. It says "extra protections" apply and that every message is checked before Copilot answers.
Two things in the same text are worth reading slowly. First, "Humans may review your conversations in Copilot for safety, legal, product improvement, or troubleshooting." That applies to everyone, teenagers included. Second, the page on the older version of the app said under-18 chats are not used to train AI models, but the page for the August 2026 app does not repeat that line. I would not promise your teenager either way.
What can I switch off at home?
Microsoft says a parent "may be able to" block the Copilot app or its website, set times of day when it can be used, and choose which features stay on, including image generation, voice mode, group chat and AI companions. Some controls need your child in your family group.
In practice, blocking is fiddly. Microsoft's block-apps page says to block the Copilot app under the Windows tab and then add copilot.microsoft.com as a blocked site, and that you repeat the process for each app, device and family member. Only the family organizer can do it.
There is also a newer switch. Microsoft's AI controls let the organizer turn supported AI features on or off per family member. Copilot, Outlook and Teams are covered now, and Word, Excel, PowerPoint and others are listed as coming soon. The page warns that a change can take a few hours to reach your child's devices.
None of those pages says a parent can read a child's Copilot chats. They describe blocking, timing and feature choices. If you want to know what your child asked, the dependable way is the one that needs no setting: ask them.
What I See in Schools
This part is my interpretation, not Microsoft's. The pattern I see most often in schools is that Copilot arrives through the account, not through a decision anyone announces. A school (in the UK, a secondary school) moves to Microsoft 365, an administrator switches Copilot Chat on for some grades, and a student opens it from the same page as their email. The student may never have thought of it as "an AI chatbot."
Microsoft's October 15, 2026 message center notice says extra content protections for users aged 13 to 17 switch on automatically for Copilot Chat, "with no additional controls," and that responses on sensitive topics "may be more limited." It also tells schools to make sure students have the right age group recorded. So a protection exists, and it depends on the school having done that bit of housekeeping. A parent cannot see that from the outside. A school can tell you in a sentence.
How common is this, really?
Not as common as ChatGPT. In a survey of 1,458 US teens aged 13 to 17, fielded in fall 2025, Pew Research Center found 64% of all US teens use an AI chatbot and 59% use ChatGPT, against 23% for Gemini and 20% for Meta AI. Its text says only that "fewer" use Copilot, with no figure given. That is one country, one age group, and a survey from before the August 2026 app. Pew's February 2026 report found 54% of US teens had used chatbots for help with schoolwork, which is the setting where a school-issued Copilot is most likely to appear.
What would I do at home?
Here are my suggestions, separate from what the sources say.
Tonight. Ask your child, without a lecture, which account they sign in with when they use Copilot. If the answer is the school account, you have your question for the school. If it is a personal account, open Family Safety and look at what is switched on.
You could say: "I'm not checking up on you. I'd like to know which Copilot you use for school, because it turns out the school's one and the home one have different rules."
By age.
- Under 8: no chatbot at all. Microsoft's own minimum age is 13, and an adult doing the typing is a different activity.
- 8 to 12: below Microsoft's age line. Keep it out of personal accounts, and treat a school-issued version as the school's responsibility to explain.
- 13 to 15: reasonable for homework help on a personal account if you have set the controls and agreed what it is for. Keep devices out of bedrooms at night.
- 16 to 18: more independence, same habit. Let AI help with the doing; keep the thinking with your child.
What no setting guarantees. An age-appropriate filter lowers the odds that a teenager meets something unsuitable. It does not make a chatbot right about facts, and it does not replace a person if your child is struggling. For feelings, my page on a teen who talks to an AI about feelings is a better starting point than any filter, and a pediatrician or the school's safeguarding lead comes before any chatbot.
When the rules come first. If your school or your country's law sets a higher age than 13, follow that. Microsoft says the age may be higher in some places, and the school's acceptable-use policy overrides anything on this page.
What to ask the school
Four questions are enough, and a good school will answer them in a paragraph:
- Is Microsoft Copilot Chat switched on for students, and for which grades?
- Have students' ages been recorded so the teen protections apply?
- What do you keep, and for how long, and who at the school can see it?
- What should families do about Copilot at home, so we are not setting different rules?
For a longer list, see 10 questions to ask your child's school about AI. 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 them the page on AI training for schools.
Other chatbots work differently, so you may want my pages on Gemini, ChatGPT's parental controls and what a school can see on a personal account. If you want a calm read on this kind of change each week, I write a newsletter that parents are very welcome to join.
Dan Fitzpatrick is the founder of The AI Educator, a former teacher and assistant headteacher, and the author of books on AI in education. More about Dan.
