Following disturbing reports of people committing suicide after interacting with ChatGPT, OpenAI decided to implement safety measures in hopes it would prevent that from happening again. This includes safeguards to protect younger users, like teens. But how will the AI know a user is under age? According to OpenAI, they now use age prediction with ChatGPT to determine that.
ChatGPT will employ age prediction
However, with age prediction, that makes it trickier. OpenAI says, “The model looks at a combination of behavioral and account-level signals, including how long an account has existed, typical times of day when someone is active, usage patterns over time, and a user’s stated age.”
So, what happens if a user claims to be over 18 but the age prediction thinks they’re not? OpenAI says that they can verify themselves with a selfie through Persona, an identity-verification service. The age prediction system will also automatically apply additional protections until the user verifies themselves.
This includes protection against graphic violence or gory content, viral challenges that are considered risky or harmful, sexual, romantic, or violent role-play, and more. Seeing as how OpenAI plans to introduce an “adult mode” to ChatGPT , this could come in handy. But the bigger question is, will it be effective?
We’ve seen it fail before
Age prediction models aren’t exactly new. In fact, back in 2025, Google attempted to use AI to predict the user’s age. It was well-intentioned, although user reports claim that it wasn’t that accurate. Users who were considered adults were finding themselves locked out of their accounts .
While Google also has a way for users to prove their age , the point is that it’s a disruptive process. Imagine trying to get work done only to be locked out because the AI did a bad job at predicting your age. Will OpenAI surpass Google in its success? That remains to be seen, but hopefully the company has learned from Google’s mistakes.
