OpenAI has started rolling out a new age prediction system on ChatGPT aimed at better identifying users who may be under the age of 18 and applying extra safety measures to their accounts. The company says the update is being released worldwide, with users in the European Union expected to receive it in the coming weeks.
In a statement posted on X on Tuesday, formerly Twitter, OpenAI said the feature is designed to improve online safety for younger users while still allowing adults full access to the platform. According to the company, the system helps ensure that minors receive added protections, even in cases where age information was not clearly provided or may be incorrect.
The company described the move as part of its wider efforts to strengthen protections for teens online. “Young people deserve technology that both expands opportunity and protects their well-being,” OpenAI said, pointing to its Teen Safety Blueprint and Under-18 Principles for Model Behavior.
OpenAI explained that teens who already state they are under 18 during signup are given additional safeguards meant to limit exposure to sensitive or harmful content. The newly introduced system goes further by identifying accounts where age details are missing or may not be accurate.
The company said age prediction is meant to strike a balance. It is designed to make sure adults are treated as adults and can use the tools as they wish within safety limits, while younger users are given stronger protections.
According to OpenAI, ChatGPT now uses “an age prediction model to help estimate whether an account likely belongs to someone under 18”. The model relies on both behavioral and account-level signals. These include how long an account has been active, usual times of use, long-term usage patterns, and any age information a user has shared.
“Deploying age prediction helps us learn which signals improve accuracy,” OpenAI said, adding that the model would be continuously refine[d] over time.
When the system believes an account may belong to a minor, ChatGPT automatically applies extra safeguards. These are meant to reduce exposure to material that could be harmful or unsuitable for younger users.
The restricted content includes “graphic violence or gory content”, “viral challenges that could encourage risky or harmful behavior in minors”, and “sexual, romantic, or violent role play”. Other areas affected include depictions of self-harm, as well as content that promotes extreme beauty standards, unhealthy dieting, or body shaming.
OpenAI said these decisions are guided by expert advice and academic research on child development. The company noted “known teen differences in risk perception, impulse control, peer influence, and emotional regulation” as key factors behind the added protections.
The company also acknowledged that the system will not always be accurate. In cases where adults are mistakenly placed in the teen experience, they are given a way to confirm their age and regain full access.
“Users who are incorrectly placed in the under-18 experience will always have a fast, simple way to confirm their age and restore their full access,” OpenAI said.
Age confirmation is done through a selfie using Persona, a secure identity verification service. Users can also check whether safeguards are active on their account by visiting “Settings > Account”.
OpenAI added that if it does not have enough information to confidently determine a user’s age, it will choose a safer experience by default.
The company also said parents can further adjust how their teenagers use ChatGPT through parental controls. These options include setting quiet hours when ChatGPT cannot be used, managing features like memory or model training, and receiving alerts if signs of serious distress are detected.
OpenAI said the rollout builds on plans it had previously announced and that it is now sharing more details as the feature becomes active. The company framed the update as part of a broader effort to improve safety for young people online, while continuing to improve how its systems tell the difference between teen and adult users.
