OpenAI is rolling out age prediction in ChatGPT to identify when an account likely belongs to someone under 18, so the system can apply the right experience and safeguards for teens. The model analyzes behavioral patterns like usage times, how long the account has been active, and the age users entered at signup. When someone gets flagged as a minor, ChatGPT automatically enables safety features that block, among other things, graphic violence, sexual roleplay, depictions of self-harm, and content about extreme beauty standards.

The move follows OpenAI's announcement that adults will get access to some of this previously restricted content, making age verification a necessary first step. It also comes after cases of teenagers developing dangerous dependencies on AI chatbots, some with fatal outcomes.

Adults who get incorrectly flagged as minors can verify their age by taking a selfie through the Persona service. Parents get additional controls, including rest periods and notifications when the system detects signs of acute distress. The feature launches in the EU in the coming weeks. More details on OpenAI's help page.