OpenAI tightens ChatGPT for teens with new study tools and parental controls

OpenAI is giving teenagers a more restricted version of ChatGPT as it tries to answer a difficult question facing schools and families: how can young people use AI without becoming too dependent on it?

The company announced ChatGPT for Teens on Tuesday, August 18, for users aged 13 to 17. The new experience introduces stronger protections around sensitive content, learning tools designed to encourage independent thinking and additional controls for parents.

The move comes as AI becomes increasingly common in education, with students using chatbots to explain difficult subjects, practise questions, research topics and complete school assignments.

OpenAI says its aim is not to keep teenagers away from AI, but to make their use of it more appropriate for their age.

“Teens are the first generation growing up with AI,” OpenAI said previously, noting that nearly nine in 10 teenage ChatGPT users use the service for learning, information, skills or productivity in a typical week.

Less homework outsourcing, more guided learning

One of the biggest changes is the greater emphasis on how students arrive at an answer.

Through Study Mode, ChatGPT can guide teenagers through problems step by step rather than simply producing a completed answer. The system can also provide quizzes and visual learning aids.

The idea is simple: if a student asks for help with a maths problem, the chatbot should be more likely to help them work through the problem than simply hand over the final answer.

OpenAI has also introduced features that can make Study Mode the default during selected study periods. Break reminders are designed to encourage teenagers to step away from the screen.

Homework reminders can also redirect students towards working through problems when the system detects attempts to bypass the learning process.

For schools and parents, that distinction matters.

The concern is not only whether students use AI. It is whether they are learning with it or simply handing their schoolwork over to it.

ChatGPT will also be less human-like

OpenAI is also placing greater distance between teenagers and the chatbot as a social companion.

Teen accounts can disable voice responses, while the system is designed not to suggest that it has feelings, emotions or consciousness.

Romantic and sexual conversations will be restricted, while additional safeguards will apply to areas including self-harm, violence, eating disorders, dangerous activities and explicit or graphic material.

The company's under-18 principles say the assistant should not become a substitute for a therapist, best friend or other real-world support.

That reflects a wider concern about young users forming unhealthy relationships with AI systems.

OpenAI has been working with outside experts, including the American Psychological Association, on guidance around young people's use of AI and mental wellbeing.

Parents get more control

The new teen experience also gives parents a bigger role.

Parents who link their accounts with their teenager's account will be able to set periods when ChatGPT cannot be used and manage selected settings.

OpenAI says parents do not automatically gain access to their teenager's conversations. Instead, the controls are intended to give families a way to manage how the service is used while preserving the teenager's privacy.

In limited high-risk situations, parents may also receive safety notifications.

OpenAI has previously said such notifications are intended for serious situations rather than ordinary conversations, political discussions or general expressions of anger.

The approach reflects a broader shift in how technology companies are dealing with minors.

Instead of relying solely on parents to police children's use of AI, companies are increasingly trying to build age-appropriate protections directly into the products.

The difficult part: knowing who is a teenager

There is, however, a practical challenge.

OpenAI says its system can estimate whether a user is under 18. If a user identifies themselves as being aged 13 to 17, they will be placed in the teen experience.

Age estimation is not perfect. That means some adults could potentially receive the more restricted experience, while systems designed to identify minors must also avoid unnecessary collection or exposure of sensitive personal information.

OpenAI has previously said it favours defaulting to the safer under-18 experience when there is uncertainty about a user's age. It has also provided mechanisms for adults to verify their age if they are incorrectly classified.

What this means for students

For teenagers, the changes could make ChatGPT feel different from the version adults use.

The emphasis is shifting towards guided learning, shorter and healthier sessions, clearer boundaries around sensitive subjects and stronger links between AI use and real-world support.

But the safeguards do not remove the need for parents, teachers and students themselves to understand the limits of AI.

A chatbot can explain a difficult concept. It can quiz a student. It can help organise ideas.

It can also make mistakes.

OpenAI's new approach therefore leaves a bigger question for classrooms and families: how do you make AI a learning partner without allowing it to become a replacement for learning itself?

That question is likely to become increasingly important as a generation of students grows up with AI available at the same time as textbooks, teachers and search engines.

About the Author

Maureen Onyango is a journalist passionate about storytelling, life coaching and spiritual lessons. She studied at the Kenya Institute of Management and enjoys telling stories that inform, inspire and empower communities.