OpenAI Will Predict User Age and Default Uncertain Cases to Teen Mode

OpenAI says it is building a long-term system to estimate whether a ChatGPT user is over or under 18, and that when the system cannot tell, it will default users into the stricter teenage experience rather than the adult one.

The company disclosed the plan in an update titled "Building towards age prediction," published on its site. "When we identify that a user is under 18, they will automatically be directed to a ChatGPT experience with age-appropriate policies," OpenAI wrote. That restricted experience includes blocking graphic sexual content and, "in rare cases of acute distress, potentially involving law enforcement to ensure safety."

OpenAI acknowledged the technical difficulty of the approach. "This isn't easy to get right, and even the most advanced systems will sometimes struggle to predict age," the company said. Its stated fallback rule is conservative: "If we are not confident about someone's age or have incomplete information, we'll take the safer route and default to the under-18 experience"—with mechanisms for adults to prove their age and unlock adult capabilities.

How the system would behave

The core design principle OpenAI articulated is that teenage users get a materially different product. "The way ChatGPT responds to a 15-year-old should look different than the way it responds to an adult," the company wrote, adding that "teens are growing up with AI, and it's on us to make sure ChatGPT meets them where they are."

When OpenAI's principles conflict, the company says it has set an explicit ordering of priorities. "When some of our principles are in conflict, we prioritize teen safety ahead of privacy and freedom," OpenAI stated, linking to a separate document on teen safety, freedom, and privacy that lays out its reasoning.

That hierarchy carries real trade-offs. An age-prediction system that errs toward the under-18 experience will place some adults into a restricted product until they verify their age, and age inference itself runs on signals the company has not detailed in this update. OpenAI framed the stakes as span-of-use rather than edge-case: it said feedback from partners and advocates is "shaping this work in ways that will matter for hundreds of millions of people."

Parental controls arriving this month

While age prediction matures, OpenAI is positioning parental controls as "the most reliable way for families to guide how ChatGPT shows up in their homes." The company reiterated that these controls will ship by the end of the month and enumerated their capabilities:

• Parents can link their account with their teen's account through a simple email invitation, with a minimum teen age of 13.
• Parents can shape how ChatGPT responds to their teen, based on teen-specific model behavior rules.
• Parents can disable individual features, including memory and chat history.
• Parents will receive notifications when the system detects their teen is in a moment of acute distress. "If we can't reach a parent in a rare emergency, we may involve law enforcement as a next step," OpenAI wrote, noting that expert input will guide the feature "to support trust between parents and teens."
• Parents can set blackout hours when a teen cannot use ChatGPT. OpenAI flagged this as "a new control we're adding."

The parental controls layer onto features OpenAI already offers or plans for all users, including in-app reminders during long sessions intended to encourage breaks.

Why it matters

The announcement lands amid intensifying scrutiny of how AI companies handle minors. Regulators and lawmakers in multiple jurisdictions have pressed for age assurance and youth-safety mechanisms on consumer platforms, and OpenAI's decision to publish its prioritization—teen safety over privacy and freedom—signals how it intends to defend those choices publicly.

The law enforcement escalation path is the most consequential element. OpenAI is stating plainly that distress detection in minors can trigger contact with police when a parent cannot be reached, a policy that puts the company in the position of making judgment calls about acute mental-health emergencies among teenagers.

OpenAI also framed the work as ongoing and externally informed. "Over the past two weeks, we've been expanding our conversations with experts, advocacy groups and policymakers who support and guide our efforts," the company wrote, and it pledged to "continue to share progress as we go."

The near-term milestone is concrete: parental controls by the end of the month, including the new blackout-hours option. The longer-term question is accuracy. An age-prediction system that defaults uncertain users to teen mode will be judged on how many adults it mistakenly restricts, how reliably it catches users under 18, and how well its distress-detection and law enforcement escalation works in the rare cases where OpenAI says those judgments will matter most.