Character AI Age Verification: A Teen's ID in a Database

When a fourteen-year-old named Sewell Setzer the third asked an app to prove he was old enough to use it, no one imagined where that data would eventually lead. On 01/07/2026, Google and Character A.I. settled a lawsuit with his family after he died by suicide. His story is the reason age verification exists on these platforms today, and the reason we need to talk about what age checks actually do to your teenager's personal information.

If your child has ever scanned their face or

If your child has ever scanned their face or uploaded an I.D. to get into an app, this touches your family directly. And I understand the instinct, you want a wall between your kid and the internet's worst corners. Demanding age verification feels like building that wall. But there's something most parents never get told. The tool meant to answer one simple question, are you old enough?, is quietly answering a very different one. So how did a safety feature turn into an identity trap?

Let's start with the technology itself, because it fails in a way almost no one expects. There are two completely different processes hiding under the same phrase. The first is facial age estimation, a system that looks at your face and guesses how old you are. The second is facial comparison, matching a selfie against an I.D. photo to confirm you're the same person. Those aren't the same thing. One estimates age. The other confirms identity. And this is the moment most parents pause and say, wait, they're not even measuring the same thing?

Here's why that matters. According to age estimation research, the technology carries a margin of error of about one and a quarter years, right at the thirteen-year-old threshold. That's the exact legal line between a child and a teen who's allowed on the platform. So a real thirteen-year-old might scan as eleven. A real eleven-year-old might scan as thirteen. And there's no software patch that fixes it, because a face simply doesn't announce its exact age.

The identity check has its own hole

Now the identity check has its own hole. When an app matches a current selfie to an I.D. photo, it's only confirming the two faces belong to the same person. It learns nothing about age. A sixteen-year-old could submit a parent's passport from eight years ago, and the system just returns a low match and moves on. The check that was supposed to protect a child ends up protecting almost nothing.

Let me address the belief that keeps a lot of us calm. Many parents assume a quick selfie is safer than uploading a document, no sensitive paperwork, no risk. People believe this because a selfie feels casual, like something you do fifty times a day. But that thirty-second selfie extracts your facial geometry, a mathematical map of your unique face. The technical process is identical whether it's a dating app or a streaming service. What almost no platform tells you is whether they delete that map once the check is done.

And that's where this gets bigger than one app. Reporting on age-check systems found that a supposedly simple age check was quietly running two hundred sixty-nine background and facial risk checks, including watchlists and political exposure flags. That's not verification anymore. That's profiling. For a compliance officer, that's a data-retention nightmare. For the rest of us, it means an age gate quietly became a file on your family.