OpenAI gives teens their own ChatGPT experience

OpenAI has rolled out ChatGPT for Teens, a new version of its AI chatbot designed specifically for users aged 13 to 17, combining learning-focused features with stronger safety protections and parental controls.

Users who state that they are between 13 and 17, or who OpenAI's systems estimate are under 18, will automatically be placed into the teen experience. The move builds on the company's existing efforts around parental controls, age prediction and its under-18 model guidelines.

At the centre of ChatGPT for Teens is a push to position the chatbot as a learning companion rather than a shortcut for schoolwork. The experience incorporates Study Mode, which uses guiding questions and step-by-step assistance to help users work through problems instead of simply giving them an answer.

Don't miss: OpenAI's first-ever influencer brand trip sparks backlash

OpenAI is also introducing "responsible homework reminders", which can identify when a teen appears to be trying to shortcut an assignment and redirect them towards Study Mode. Quizzes and learning visualisations are included to help users test their understanding, while a new Study Hours feature allows teens or their parents to set periods when Study Mode is switched on by default.

According to OpenAI, the learning features are intended primarily to support students outside the classroom, while its separate ChatGPT for Teachers product gives educators and school districts institution-managed access to its AI tools.

Alongside the launch, OpenAI is partnering with CodeAI to provide students and educators with resources aimed at building AI literacy. The partnership will focus on helping teens understand how AI works, how to direct and question it, and how to use it creatively, while keeping teachers involved in how the technology enters classrooms.

Safety, meanwhile, is being built into the teen experience by default. ChatGPT for Teens applies age-appropriate protections intended to reduce exposure to potentially harmful or developmentally inappropriate content, particularly around self-harm, violence, eating disorders, dangerous activities and explicit sexual or graphic material.

Parents with linked teen accounts can use existing tools such as Quiet Hours and manage selected settings, while receiving notifications in limited high-risk situations. OpenAI is also expanding these notifications to cover situations related to eating disorders.

The company is introducing additional features aimed at encouraging healthier relationships with AI, including break reminders and cues that reinforce that users are interacting with an AI system rather than a person.

Under OpenAI's updated under-18 model guidelines, ChatGPT should not engage teens using romantic language, encourage emotional dependence or suggest that it has feelings or consciousness. Teen users will also receive reminders before uploading potentially sensitive images, alongside teen-specific onboarding explaining the learning and safety features available to them.

Personalisation will still form part of the experience, with options such as accent colours and voice variations. OpenAI said these are intended to give teens some control over how ChatGPT looks and sounds without blurring the distinction between the chatbot and a human relationship.

The launch comes as AI companies face growing scrutiny over how younger users interact with conversational AI, particularly as chatbots become increasingly embedded in education and everyday digital behaviour.

OpenAI said it will continue testing its safeguards against teen-specific standards and publishing findings on areas including self-harm, eating disorders, violence, age-restricted products and services, and sexual content as it develops the experience further.

The teen-focused protections also follow OpenAI's broader push to introduce safety features aimed at connecting users with real-world support. Earlier this year, the company began rolling out Trusted Contact, an optional ChatGPT feature that allows adult users to nominate a person who may be notified if the system detects signs of serious self-harm-related risk during conversations.

The feature allows users to select one trusted person, such as a friend, family member or caregiver. The nominated contact must accept an invitation before the feature is activated.

Related articles: 
OpenAI shuts down Sora, reportedly ending Disney partnership talks   
OpenAI pushes back against order to hand over millions of ChatGPT convos   
PayPal partners OpenAI to drive agentic commerce in ChatGPT