OpenAI rolls out age prediction for ChatGPT accounts to automatically estimate your age

OpenAI is rolling out age prediction for ChatGPT consumer plans to better identify accounts likely belonging to teenagers and apply safeguards tailored to minors. The age prediction model uses a mix of behavioral and account-level signals, such as how long an account has existed, common usage times, stated user age, and long-term usage patterns to help estimate a user's age.

For accounts suspected to be used by those under 18, ChatGPT automatically activates expanded protections. These measures aim to reduce exposure to sensitive content, including graphic violence, viral challenges, sexually explicit material, depictions encouraging self-harm, promotion of extreme body standards, unhealthy dieting, or body-shaming. While teens who state their age as under 18 at sign-up already receive enhanced content controls, age prediction extends protections to more users who might not self-identify.

Building on these safeguards, the approach is guided by expert advice and academic research regarding differences in teens’ risk perception, impulse control, and peer influence. In addition to automated measures, parents can further tailor their teen's ChatGPT experience via parental controls.

Should an adult be wrongly placed under these restrictions, they can swiftly restore full access with a selfie-based age verification through Persona, a secure identity service. OpenAI also uses feedback from this system to make ongoing improvements to the model’s accuracy.