Is ChatGPT safe for kids and teens? What ChatGPT for Teens changes

Key points of this article

• ChatGPT requires users to be at least 13 years old, and those under 18 need a parent's or guardian's permission.
• Since August 18, 2026, accounts belonging to users who state that they are 13 to 17 years old, or whom OpenAI's age-prediction model estimates to be under 18, are automatically placed in ChatGPT for Teens.
• ChatGPT for Teens reduces graphic, sexual and self-harm content by default, adds break reminders, and steers teens to Study mode when they seem to be shortcutting homework.
• Parents can link accounts to set quiet hours, switch off features and get rare safety alerts, but they cannot read their teen's conversations.
• The protections only work in the teen's own official ChatGPT account. A shared adult account, an unofficial "unblocked" site, or a fake ChatGPT app or extension bypasses them.

Imagine your teenager struggling with a math problem late in the evening. Instead of asking you for help, they open ChatGPT. Within seconds, they have an explanation and perhaps even the solution. But what happens when the questions become more personal or when the chatbot gets something wrong? OpenAI has introduced stronger protections for teenagers, but no technology can eliminate every risk. Understanding these safeguards and their limitations can help families make more informed decisions about AI use.

Is ChatGPT safe for kids?

ChatGPT isn't built for children under 13, and its terms don't allow them to use it. Teens aged 13 to 17 get to use ChatGPT for Teens, which turns on stricter content protections by default and lets parents link accounts to set limits. It's safer than before, but filters can be bypassed, and parents can't read chats.

A Pew Research Center survey of 1,458 U.S. teens, published in February 2026, found that 54% had used AI chatbots for help with schoolwork and that 12% had used them for emotional support or advice.

These are very different situations. An inaccurate answer to a history question might affect an assignment and possibly a school grade, while inappropriate personal advice could have even more serious consequences. Some families have even sued OpenAI, alleging that chatbot interactions contributed to harm involving teenagers. The question is therefore not simply whether ChatGPT is safe but how teenagers use it and how well they understand its limitations.

Want to know more about ChatGPT's general security and privacy risks? Explore ESET's complete 2026 guide to ChatGPT safety.

Read on

Expert insights

Kamila Ryšánková, a child psychologist, explains: “AI can be very helpful for children who struggle with learning or have low self-esteem. It offers them a safe space where they can ask questions, make mistakes, and learn at their own pace without fear of being ridiculed or judged by others. It can explain material in simpler terms, provide examples, and repeat information as many times as necessary until they understand it. As a result, children can develop greater confidence, independence, and motivation to learn.

However, it is important that AI does not replace the support of parents, teachers, and psychologists. Children also need understanding, encouragement, and reassurance that mistakes are a natural part of learning. The goal should be for AI to help children become less afraid of asking questions and develop greater trust in their own abilities. Interactions with peers are also essential, as they help children develop social skills, learn to cooperate, and resolve conflicts. AI cannot fully replace these experiences.

Parents and teachers should encourage children to try solving a task on their own first and only then turn to AI, for example, when they do not understand something or need further explanation. One useful technique is the rule: ‘Me first, then AI.’ The child first tries to solve the task independently and writes down their ideas. Only then do they use AI and compare its answer with their own solution. Another option is to ask AI for a hint instead of the complete answer, for example: ‘Give me a hint about how to approach this task, but do not reveal the solution.’ It is also important to praise children’s efforts, their own ideas, and their problem-solving process rather than focusing solely on correct answers. This helps children become more independent, develop critical thinking skills, and gain greater confidence in their own abilities.

The main psychological risks of using AI include excessive reliance on technology, weakened independent thinking, and reduced communication with other people. Children may begin turning to AI even with personal problems instead of confiding in their parents, friends, or other trusted individuals. This can be particularly concerning for children who feel lonely or have low self-esteem.

AI does not manipulate children intentionally, but it can still influence them to some extent. Children may trust its answers because it sounds confident and friendly, even when the information it provides is inaccurate or inappropriate. They may also adopt opinions without critically evaluating them or develop an overly strong emotional attachment to AI. Another risk is that AI may fail to recognize the seriousness of a child's problem and offer advice that is unsuitable for the situation. Therefore, it is important to teach children to evaluate AI-generated responses critically, verify information, and understand that they can turn to a trusted adult when dealing with sensitive personal issues.

One warning sign may be a situation in which an adolescent spends so much time interacting with AI that they neglect school, sleep, hobbies, or contact with friends and family. It may also be concerning if they begin prioritizing AI over real-life relationships and share all their problems with it because they feel that other people do not understand them. Other warning signs may include significant mood changes, irritability when access to AI is restricted, increasing uncertainty when making independent decisions, or allowing AI-generated responses to influence important personal decisions.

The amount of time spent using AI is not necessarily a problem in itself. What matters is whether AI negatively affects the adolescent's everyday life, psychological well-being, and interpersonal relationships. If significant or persistent changes in behavior, mood, or relationships occur, it is advisable to talk openly with the adolescent without judging them and, if necessary, seek professional help.”

What is ChatGPT for Teens?

ChatGPT for Teens is the version of ChatGPT that OpenAI launched on August 18, 2026, for users aged 13 to 17. It switches on age-appropriate content protections by default, adds learning tools such as Study mode and homework reminders, and works with optional parental controls. Rather than being a separate chatbot, it combines existing safeguards with new features. Some, including parental controls and Study mode, were already available before its launch.

What does ChatGPT for Teens offer?

These protections build on OpenAI's Under-18 Principles and Teen Safety Blueprint. One important principle is that ChatGPT should not encourage emotional dependence or imply that it has human feelings. A chatbot may respond like a person, but it is not one. Teen safeguards apply automatically to eligible accounts. Additional parental controls require parents and teenagers to link their accounts.

How does ChatGPT know a user is a teen?

ChatGPT puts an account in ChatGPT for Teens if the user states an age from 13 to 17 or if OpenAI's age-prediction model estimates the account likely belongs to someone under 18. The model weighs signals such as account age, typical activity times, usage patterns and stated age. When unsure, it applies the teen experience.

OpenAI introduced age prediction in January 2026, with an EU rollout that began in August. Adults incorrectly classified as teenagers can verify their age using a selfie through OpenAI's identity-verification partner Persona.

It is important to keep in mind that the protections follow the account, not the device. A teenager using a parent's logged-in adult account on their device will not receive the safeguards associated with their own account.

Check your teen's account in Settings > Account to see whether teen safeguards are applied, and make sure your child does not use your device with your account logged in.

What can parents see and control in ChatGPT?

Parents can link their ChatGPT account to their teen's to change settings, set quiet hours, and receive safety notifications in rare high-risk situations. They can't read their teen's conversations, see chat history, or monitor activity in real time. Either side can unlink, and the parent is notified if the teen does.

OpenAI says notifications may be sent when trained reviewers identify serious concerns, including signs of self-harm or disordered eating. Notifications can also follow certain account deactivations related to violence. These alerts share limited information, not conversation transcripts, and are not a substitute for professional care.

How to set up ChatGPT parental controls

1. Open ChatGPT on the web or in the app, and go to Settings > Parental controls.
2. Select "+ Add family member," and send your teen an invitation. Teens can also invite a parent.
3. Ask your teen to accept the invitation. Each teen's account can be linked to one parent or guardian at a time.
4. Select your teen under Family members, and adjust their settings.
5. Choose how you want to receive safety notifications: email, text message or push alert.

Recommended settings

• Keep "Reduce sensitive content" enabled, and consider setting quiet hours from 10 p.m. to 7 a.m. on school nights. For homework, turn on Study mode for eligible new conversations or schedule Study Hours.
• For younger teenagers, consider disabling memory. Turning off "Improve the model for everyone" prevents their conversations from being used for model training.
• Review voice and image features together, and keep safety notifications enabled.
• Discuss the settings with your teenager so they understand their purpose.

What risks remain with ChatGPT for Teens?

ChatGPT for Teens lowers some risks but doesn't remove them. OpenAI states that its guardrails aren't foolproof and can be bypassed. The main remaining risks are wrong answers, emotional over-reliance, oversharing personal data and use outside the protected teen account.

Imagine, for example, a teenager asking ChatGPT about a concerning medical symptom or uploading a school document containing classmates' personal details. Neither situation is necessarily prevented by content filters.

If an AI conversation becomes upsetting, teenagers should know they can stop and speak to someone they trust.

Need immediate support?

If your teen is struggling or talks about harming themselves, call or text 988 to reach the 988 Suicide & Crisis Lifeline (U.S.). If there is immediate danger, call 911.

In the EU, children and teens can call 116 111 for free, confidential support in participating countries. In case of immediate danger, call 112.

ChatGPT is just one of many AI chatbots children may encounter. Discover the broader risks in our article.

Find out more

How to spot fake ChatGPT apps, extensions and "unblocked" sites

ChatGPT for Teens protections and parental controls only work inside OpenAI's official ChatGPT apps and website. Copycat apps, AI browser extensions and "ChatGPT unblocked" sites used to get around school filters sit outside those protections, and some are built to steal logins or chat history.

In 2024, ESET reported more than 4,000 attempts to install a malicious extension disguised as a legitimate tool. Other researchers later identified two Chrome extensions that collected ChatGPT and DeepSeek conversations and that had approximately 900,000 combined users.

For teenagers, an unofficial AI tool may look like a convenient shortcut. But bypassing school restrictions can also mean losing the safeguards provided by the official platform.

Five ways to avoid fake AI tools

1. Install ChatGPT only from the Apple App Store or Google Play, published by OpenAI.
2. Use chatgpt.com, not mirror sites or "unblocked" links shared at school.
3. Check the browser extensions on your teen's computer, and remove AI sidebars you don't recognize.
4. Turn on multifactor authentication for your teen's ChatGPT account, and use a unique password.
5. Use security software that blocks malicious websites, apps and extensions.

Help your family explore AI more safely.
Good AI habits remain essential. AI Conversation Security in ESET HOME Security Premium adds an extra layer of security when using tools like ChatGPT by checking links in AI responses and warning about phishing or malicious websites.

Is using ChatGPT for homework cheating?

It depends on the school's rules and how ChatGPT is used. Asking it to explain a concept, quiz you or check your reasoning is generally treated as studying. Submitting text or answers it wrote as your own work usually breaks academic-integrity rules.

The 2026 Pew survey found that 10% of American teenagers use chatbots to help with all or most of their schoolwork. In a separate 2025 survey, 54% considered ChatGPT acceptable for researching new topics, compared with 18% for writing essays.

ChatGPT's Study mode encourages students to work through problems using questions, explanations and knowledge checks rather than simply receiving answers. Homework reminders can also direct teenagers toward this approach.

If your teen can't explain the answer without the screen, they haven't learned it yet.

How to turn on Study mode in ChatGPT

1. Teen: Start a new chat, open the tools menu (+), and select Study and learn.
2. Parent: Go to Settings > Parental controls, select your teen, and switch on Study mode so every new chat starts in it.
3. Parent or teen: Set Study hours to switch on Study mode automatically at scheduled times.

Study hours affect eligible new chats during scheduled periods. They do not block access outside those hours or apply to existing chats or conversations within GPT or Projects.

Which AI tools are built for students?

Several AI tools now offer a tutoring feature that asks questions instead of handing over answers, including ChatGPT's Study mode, Google Gemini's Guided Learning, and Khan Academy's Khanmigo. They differ most in minimum age and in what parents can see.

Sources: OpenAI, Google, Khan Academy, Anthropic. Availability varies by region and account type.

In a September 2025 assessment, Common Sense Media rated Gemini's teen experience as “High Risk” overall. Educational features alone are therefore not a guarantee of age-appropriate interactions.

ChatGPT safety checklist for parents

1. Confirm that your teen is at least 13 years old and uses their own account, not yours.
2. Link accounts in Settings > Parental controls.
3. Keep "Reduce sensitive content" switched on.
4. Set quiet hours for school nights.
5. Switch on Study mode by default, or schedule Study hours.
6. Decide on memory, voice mode, image generation, group chats and Health feature.
7. Turn off "Improve the model for everyone" for your teen.
8. Turn on multifactor authentication, and use a unique password.
9. Check installed apps and browser extensions for ChatGPT copycats.
10. Agree to family rules: what not to share and when to ask a person instead.

Helping teenagers make the most of AI safely

ChatGPT for Teens gives families more ways to shape how younger users interact with AI. But stronger safeguards cannot guarantee accurate answers, appropriate conversations or protection from every online threat.

The most useful approach combines technology with communication. Explore the settings together, discuss what your teenager uses AI for, and agree on boundaries. The goal is not to prevent young people from using technology but to help them recognize when it supports learning, when information needs checking and when a real person is the better source of advice.

Frequently asked questions

Is there a kids version of ChatGPT?

No. There is no version of ChatGPT for children under 13. ChatGPT for Teens covers ages 13 to 17. For younger children, Google allows Gemini on supervised Family Link accounts in some countries, and Khan Academy's Khanmigo is available to under-18s when a parent or school grants access.

How old do you have to be to use ChatGPT?

You must be at least 13 years old to use ChatGPT, and users under 18 need a parent's or guardian's permission. Accounts with a stated age of 13 to 17, or those that OpenAI's age-prediction model flags as likely belonging to users under 18, are automatically placed in ChatGPT for Teens.

Can parents see their teen's ChatGPT conversations?

No. ChatGPT's parental controls don't give parents access to conversations, chat history or real-time activity. In rare situations, when OpenAI's systems and trained reviewers detect signs of serious risk, a linked parent may receive a safety notification with only the information needed to support the teen.

Can teens get around ChatGPT for Teens?

Sometimes. OpenAI says its guardrails aren't foolproof. The protections are tied to the teen's own ChatGPT account, so they don't apply if a teen uses a parent's adult account, an unofficial ChatGPT site or app, or a third-party tool built on OpenAI's models.

Why is ChatGPT asking me to verify my age?

ChatGPT asks for age verification when its age-prediction model estimates an account may belong to someone under 18. An adult placed in the teen experience by mistake can restore full access by verifying with a selfie through Persona, OpenAI's identity-verification partner, from Settings > Account.

What happens to parental controls when my teen turns 18?

When ChatGPT identifies a linked teen account as 18 or older, the account can move out of ChatGPT for Teens. That can remove teen protections and end the parental-controls link. Some protections, such as reducing sensitive content, can be switched back on.