OpenAI says if it can't verify your age, you're a minor

OpenAI is rolling out a significant policy shift that defaults to treating unverified users as minors, a move the company frames as prioritizing teen safety over privacy or unfettered access. The company is building a long-term system to guess whether a user is over or under 18. When the system isn’t confident—which, as any parent of a teenager will tell you, is a frequent problem for age-gating technology—it will automatically route the user into a restricted, under-18 experience. That means blocking graphic sexual content and, in what the company calls rare cases of acute distress, potentially looping in law enforcement.

This isn’t just a content filter. The under-18 experience will fundamentally alter how ChatGPT responds to a 15-year-old versus an adult. OpenAI says the tone and behavior of the model will shift, though they’re light on specifics about what that looks like in practice. The bigger bet here is that the company can even build a reliable age-detection system at all. They admit it’s hard, and even advanced systems will stumble. The escape hatch for adults who get miscategorized is a process to prove their age and unlock the full product.

While that age-guessing infrastructure gets built, OpenAI is shipping parental controls by the end of the month. Parents with kids 13 and up can link accounts, disable features like memory and chat history, and set blackout hours—a direct nod to the screen-time anxiety that dominates family dinner tables. They’ll also get notifications if the system detects their teen is in a moment of acute distress. If a parent can’t be reached, OpenAI reserves the right to call the cops. Expert input, the company says, will guide that feature to balance trust and safety.

The whole package is a tightrope walk. OpenAI is openly declaring that when its principles clash, teen safety wins. That sounds noble on paper, but the mechanics of age detection and the decision to involve law enforcement in edge cases will raise immediate questions about overreach and false positives. As the company thanks its partners and advocates for pushing them, the reality is that this system will be judged not on its intentions, but on the kids who slip through the cracks and the adults who get locked out.