OpenAI has officially announced ChatGPT for Teens, a specialized new experience tailored for users between the ages of 13 and 17. The platform is designed around promoting learning, encouraging healthier usage habits, and instituting significantly stronger safety protections.
The company notes that nearly nine in 10 teens who use ChatGPT turn to the platform to learn, find information, develop skills, or get things done during a typical week.
Stricter Guardrails and Age Prediction
Stricter guardrails are now being put around accounts for users under 18. This is particularly true when it comes to sensitive topics such as graphic violence, dangerous activities, unhealthy dieting, extreme beauty standards, and sexual or romantic roleplay.
To enforce this boundary, OpenAI is deploying new age-prediction technology designed to keep kids under 18 out of the standard version of ChatGPT. According to the announcement, if the system cannot confidently estimate someone’s age, ChatGPT will automatically default to the under-18 experience.
“We prioritize safety ahead of privacy and freedom for teens; this is a new and powerful technology, and we believe minors need significant protection,” wrote OpenAI CEO Sam Altman in a blog post outlining the changes.
Rethinking AI in Education
At its core, the teen version of ChatGPT wants to teach, not just hand over the answer. What is most notable is how OpenAI is approaching homework with this new system.
Furthermore, the company is actively pushing that educational idea outside of the ChatGPT application, too.
However, the deeply interactive nature of the technology presents unique psychological challenges. When an AI can talk back instantly, remember the context of a conversation, and respond at 2 AM when nobody else is awake, it can be easy to forget that you’re still talking to software.
To combat this phenomenon, the company is adding a few smaller reminders to reinforce the vital distinction between artificial intelligence and a human being.
Unresolved Safety Questions
Despite these extensive new features, the rollout leaves lingering questions about how comfortable parents will actually feel with all of this.
While strict boundaries are actively being set by the company, safety measures remain imperfect. WIRED’s past testing of OpenAI’s GPT-5 model showed that certain aspects of potential guardrails could still be circumvented with basic commands in the custom instructions.