Is ChatGPT safe for my teenager? The rules I use with under-18s in class

Your teen is probably already using ChatGPT, whether that's for homework, for questions they'd rather not ask you, or just to see what it says. So the question parents ask me more than any other is whether that's safe. I teach teenagers live on Zoom and AI tools come up in my classes all the time, so I've had to work out a proper answer for my own students as well.

The short answer is that it depends on four things: how old your teen is, which tool they're using, how the account is set up, and the habits they build around it. A 16-year-old using ChatGPT with a parent's permission and parental controls switched on is in a very different position from a 12-year-old on an app meant for adults.

In this post I'll go through what the companies themselves say about age, what parental permission involves in practice, and the rules I use with under-18s in my classes.

What the age rules say (as of September 2026)

I read each company's terms and help pages in September 2026. Companies update these, so please click through and check the links yourself before you decide anything.

OpenAI (ChatGPT). OpenAI's Terms of Use (effective January 1, 2026) say you must be at least 13, or the minimum age required in your country to consent to use the services. If you're under 18, you must have your parent or legal guardian's permission. The Europe Terms of Use, which apply in the UK, EEA and Switzerland (updated January 16, 2026), say the same, and add that under-18s should ask a parent or guardian to read the terms with them. OpenAI's help page, Is ChatGPT safe for all ages?, says ChatGPT is not meant for children under 13, that ages 13 to 18 need parental consent, and that it "may produce output that is not appropriate for all audiences or all ages."

Anthropic (Claude). Anthropic's Consumer Terms of Service (effective October 8, 2025) say you must be at least 18, or the minimum age required to consent in your location, whichever is higher. That means a personal Claude account isn't meant for anyone under 18.

Google (Gemini). Google's help page What you need to sign in to Gemini Apps says you must be 13 (or the applicable age in your country) or over to use Gemini with a personal or school Google Account, and 18 or over with a work account. For children under 13, Google's page Guide your child's Gemini Apps experience says a parent must enable access through a supervised Family Link account. It also says this isn't available for supervised accounts in the European Economic Area, Switzerland and the United Kingdom. The age at which someone can manage their own Google Account is 13 in most countries, but Google lists exceptions by country. I couldn't find a separate parental-consent rule for 13 to 17 year olds on Google's Gemini help pages, so I won't claim there is one, or that there isn't. If it matters for your family, check with Google directly.

If your teen uses an AI tool through a school account, the school's agreement with the company may set different rules, so it's worth asking the school.

What parental permission means in practice

OpenAI's wording is worth reading closely. It says a teen must have your permission, which puts the responsibility on the family. In my view, that permission should come out of a conversation between you and your teen, rather than a box they tick on their own at midnight.

OpenAI also offers parental controls. According to its help page, a parent can link their account to their teen's account and manage settings such as a "Reduce sensitive content" safeguard (on by default), voice mode, image generation, memory, and "quiet hours" when the teen can't use ChatGPT. Parents may get safety notifications in limited, serious situations.

It's also worth knowing what the controls don't do. OpenAI's page says linking accounts doesn't let a parent read their teen's conversations or chat history. Either person can unlink at any time, and the parent is notified if that happens. And safety notifications "are not real-time monitoring." The controls are useful, but you'll still need to talk to your teen about how they use it.

The risks I watch for

News stories tend to focus on the most extreme cases. The risks I think about as a teacher are more ordinary, and they're the ones I'd watch for at home too.

Confident wrong answers. OpenAI's own terms warn that output "may not always be accurate." A wrong answer is written just as fluently as a right one, and teens tend to trust whatever sounds sure of itself.

Oversharing. Teens will type things into a chat box that they would never post publicly: their full name, their school, a friend's problem, a photo.

Letting it do the thinking. Cheating is one worry. The bigger one for me is a teen who stops working through hard problems, because that effort is where most of the learning happens.

Treating it like a friend. Google's guidance for parents makes this point well. Help your child understand that it's an AI tool, and not a person to confide in.

Inappropriate content. The companies use filters, but as Google puts it, they're "not perfect."

The rules I use with under-18s in class

These are the rules I follow and ask my students to follow. I keep the list short, because teenagers don't read long lists of rules, and to be fair, most adults don't either.

1. The right tool for the age. I only ask students to create accounts on tools whose terms allow their age, and only with a parent's permission. If a tool is for adults only, students don't sign up. If it's useful, I demo it from my own screen instead.
2. Parents hear first. Before a course starts, families are told which tools we'll use and why.
3. No personal information in prompts. That means no full names, addresses, school names, photos of people, or anything about friends and family.
4. AI output is a draft, never a source. If an AI tool tells you a fact, check it somewhere reliable before you use it.
5. Say when AI helped. Students note which parts of their work were AI-assisted, so it's always clear what they did themselves.
6. You have to be able to explain it. If you can't explain something in your own words, it isn't finished.
7. If something feels off, stop and tell an adult. Reporting something strange or upsetting never gets anyone in trouble.

In live classes we also use AI together, sharing screens when it helps, rather than each student chatting privately with a tool during the session.

What I'd suggest at home

You don't need to know much about technology to set good habits. These are the ones I'd start with.

• Read the tool's terms with your teen, as OpenAI's Europe terms suggest.
• If you allow ChatGPT, set up the parental controls together, so it feels like a decision you made with your teen rather than a way of checking up on them.
• At least at first, keep AI use to shared spaces like the kitchen table.
• Once a week, ask your teen to show you something they made or learned with AI.
• Write down your family's version of the rules above and put it somewhere you'll both see it.

What you're aiming for is a teen who uses these tools sensibly when you aren't there to check.

So, is it safe?

For a 13 to 17 year old using a tool whose terms allow their age, with your permission, sensible settings and good habits, I think AI can be used reasonably safely, and it can be very useful. For under-13s, and for tools that are 18+ only, the companies' own rules answer the question for you.

Beyond that, I'd worry less about which tool it is and more about two things: whether a trusted adult is involved, and whether your teen has learned to question what the tool tells them.

If you'd like to see exactly how I handle AI tools, accounts and privacy in my classes, it's all on my safeguarding page.