The tool relies on analyzing account usage patterns to estimate whether the user is under 18 years old. Indicators include method and time of use, in addition to the account creation period.
If the system suspects that the user is a minor, ChatGPT automatically activates protection measures to limit their exposure to sensitive content, such as sexual or violent material, self-harm scenes, or any content that encourages dangerous and unhealthy behaviors.
OpenAI explained in a post on its official website that this approach is based on expert opinions and academic studies in child development, taking into account the differences between adolescents in risk perception, impulse control, peer influence, and emotional regulation.
A user who has been wrongly classified as under 18 years of age can confirm their real age by uploading a government-issued ID, such as a driver's license or passport, or by sharing a selfie with the Persona identity verification service, which uses artificial intelligence to determine age.
This feature comes ahead of the expected launch of "Adult Mode" in the coming weeks, which will allow users to create explicit and violent content using ChatGPT's image, text, and video tools, and discuss sensitive topics that are currently restricted, while non-consensual content, deepfakes, and illegal acts will continue to be banned.
The release date for "Adult Mode" has not yet been determined, but Fidji Simo, the company's head of applications, stated last December that it is expected to launch in the first quarter of 2026.
