ChatGPT’s New Age Detection System Includes ‘Selfie’ Verification

OpenAI has officially launched a new system that aims to predict your age by watching how you use ChatGPT.

Announced on Tuesday, the AI giant is rolling out an “age prediction” model across its ChatGPT consumer plans. Instead of just taking your word for it, the system now looks for digital clues to determine if an account belongs to a minor.

According to the company, “When the age prediction model estimates that an account may belong to someone under 18, ChatGPT automatically applies additional protections⁠ designed to reduce exposure to sensitive content.” 

These safeguards are designed to block the generation of content involving graphic violence, sexual or violent role-play, depictions of self-harm, and material promoting extreme beauty standards or unhealthy dieting.

The ‘teen safety’ push

This isn’t happening in a vacuum. OpenAI, like many tech firms, is under growing legal and regulatory pressure to better protect younger users. The company is named in lawsuits, including one reported by CNBC that centers on a teenage boy’s death by suicide allegedly linked to ChatGPT interactions. 

The Federal Trade Commission is also probing how AI chatbots affect children. 

OpenAI frames the new system as a way to ultimately give adults more freedom. “This also enables us to treat adults like adults and use our tools in the way that they want, within the bounds of safety,” the company’s announcement explained. The idea is that by confidently identifying minors, the platform can apply stricter rules where needed without limiting all users.

What if the AI gets it wrong?

The system won’t be perfect. Adults who find themselves incorrectly flagged as teens will have a path to restore full access, but it involves an extra step. They’ll need to verify their age through Persona, a third-party identity service, by submitting a selfie.

Users can check whether safeguards have been applied to their account in Settings > Account, where they can start the verification process with a selfie.

Alongside age prediction, OpenAI says parents can further customize their teens’ ChatGPT experience using parental controls. These tools allow parents to set quiet hours, manage features like memory or model training, and receive notifications if signs of acute distress are detected.

The company says these controls are meant to give families greater visibility and choice as teens use AI tools more frequently at school and at home.

The rollout is global but will hit the European Union in the coming weeks to comply with regional digital safety laws. OpenAI says it will keep refining the model’s accuracy and that its broader work on teen safety, developed with groups like the American Psychological Association, is ongoing.

Also read: AI trends for 2026 to learn more about where Microsoft says AI is headed next.