ChatGPT for Teens: Parent's Guide

AI Summary

What this is: OpenAI launched a separate teen version of ChatGPT on 18 August 2026. This covers what changed, what parents can control, and what the research actually says about teens and AI.

Who it’s for: Parents and guardians of 13 to 17 year olds, and teachers who get asked about this.

The key idea: The safety changes are real, and parental controls do not let you read your teen’s conversations. Quiet Hours is the setting most worth knowing about.

Skip if: You want general ChatGPT guidance. Start with how to use ChatGPT instead.

OpenAI began rolling out ChatGPT for Teens on 18 August 2026, a separate version of the app for 13 to 17 year olds with different rules about what the model will say. It is available globally on free and paid personal plans.

The timing is the uncomfortable part. Teenagers have been using the ordinary version for years.

Table of Contents

What actually changed?

Teens land in the teen version either by telling ChatGPT their age or by the system estimating that they are under 18. Once there, a set of protections is on by default, covering self-harm, violence, eating disorders, dangerous activities, and sexual or graphic content.

Three behavioural changes matter more than the content filters, and they are aimed squarely at the companionship problem. In the teen version the model will not use romantic language, will not encourage a teen to depend on it emotionally, and will not claim to have feelings or consciousness.

On the schoolwork side there is Study Mode, which answers with guided questions rather than finished answers, along with quizzes and a prompt that notices when a teen is trying to skip the work and steers them back toward the steps.

What can parents control?

Controls require a linked account, meaning the teen has to connect theirs to yours. Once linked, you can:

Set Quiet Hours, blocking access during chosen times. This is the one worth setting tonight if you set nothing else, because it addresses the 2am problem rather than the content problem.
Set Study Hours, so new chats open in Study Mode during homework time without blocking the app entirely.
Turn off voice mode.
Turn off memory, so ChatGPT stops saving details and drawing on them in later conversations.
Remove image generation.
Receive alerts in a narrow set of high-risk situations, such as discussion of self-harm or suicide. Eating disorder alerts were added at launch.

Parental controls do not let you read or monitor your teen’s conversations, which surprises a lot of parents. Alerts fire only in a narrow set of dangerous moments. There’s no transcript to read.

Whether that is the right design is a real debate. It does mean the tool is not a surveillance product, and a teen who knows their messages are private may raise something they would otherwise hide.

How many teens are doing this?

Common Sense Media surveyed 1,060 US teens aged 13 to 17 in a nationally representative study published in July 2025. The findings are more mixed than the headline number suggests.

Finding	Share of teens
Have used an AI companion at least once	72%
Use one at least a few times a month	52%
Use one daily or several times a week	34%
Have discussed something serious with AI instead of a person	About 1 in 3
Find AI conversations less satisfying than human ones	About 2 in 3
Spend more time with real friends than with AI	80%

Most teens are not replacing their friendships, and most find the machine worse company than a person. The number that should hold your attention is the middle one: roughly a third have taken something serious to a chatbot instead of a human being.

Common Sense Media’s own recommendation, published before this teen version existed, was that nobody under 18 should use AI companions until stronger protections are in place.

Why did OpenAI do this now?

Pressure, from several directions at once. Researchers posing as struggling teens found that the older version would warn them off risky behaviour and then supply detailed information anyway. Florida became the first state to sue OpenAI over how it described its protections for minors.

None of that makes the changes less real. It does mean this is a response, not a plan built ahead of the problem.

Does age checking actually work?

Not reliably, and this is the part most likely to fail. Placement runs on self-reported birthdays and the system’s own estimate of a user’s age. A teenager who wants the adult version can generally get it.

Treat the teen version as a meaningful default rather than a barrier. It changes what happens when a 15 year old uses ChatGPT the ordinary way, without stopping a determined teen from finding a way around it. No announcement has claimed otherwise.

What should a parent do?

Link the account and set Quiet Hours. Overnight access is where the companionship pattern takes hold, and this is a five-minute change.
Ask what they use it for before changing settings. Homework help, boredom and curiosity are the most common answers, and they call for different responses than emotional reliance does.
Turn off memory if the relationship worries you. Memory is what makes a chatbot feel like it knows you.
Do not assume you can read the conversations. You cannot, so the conversation with your teen remains the actual tool.

What AI can’t do here

A content filter cannot supply the thing a teenager is looking for at 2am. It can only decline to imitate it badly.
Age estimation cannot confirm who is actually typing, so every protection here rests on an assumption the system cannot verify.
No setting substitutes for knowing whether your kid is lonely. Most teens still prefer people. That’s good news, and it means the follow-up is a conversation with your kid, not a setting.

Know What Your Kids Are Using

Human curated and edited, for beginners who want a finger on the pulse without the hype.

Free forever. Unsubscribe anytime.

Common questions about ChatGPT for Teens

Can I read my teen’s ChatGPT conversations?
No. Parental controls do not include reading or monitoring chats. You receive alerts only in a narrow set of high-risk situations.

Is it free?
Yes. It is rolling out globally on both free and paid personal ChatGPT plans.

How does ChatGPT know my child is a teenager?
Through a self-reported birthday or the system’s own age estimate. Neither is reliable, and a teen who wants the adult version can usually reach it.

What are Quiet Hours?
A parental setting that blocks ChatGPT during times you choose, such as overnight. It is the single most useful control for the late-night use pattern.

What is Study Mode?
A mode that guides a teen toward an answer with questions and prompts, instead of just handing over the finished answer. Study Hours can make new chats open that way automatically.

Does this apply to Claude or Gemini?
No. This is an OpenAI product change. Other assistants have their own policies for younger users, and they differ.

Should my teen use AI at all?
Common Sense Media advised against under-18s using AI companions until protections improved. For schoolwork and general questions the calculation is different, and Study Mode is built for exactly that.

Sources

OpenAI, Introducing ChatGPT for Teens (primary source)
OpenAI Help Center, Parental controls on ChatGPT FAQ
Common Sense Media, Nearly 3 in 4 teens have used AI companions (survey of 1,060 US teens, July 2025)
TechCrunch, 72% of US teens have used AI companions

Google Wants a Dead Airline’s Data

How to Turn Off AI Everywhere