Meta scans faces to verify teens

Meta is tightening age checks on Instagram and Facebook, and the new part is visual analysis. The company said on May 5 that it will use AI not just to read obvious account signals, but also to look at photos and videos for broad age clues. The goal is twofold — push suspected teens into Teen Account protections, and catch accounts that may belong to children under 13, who are not allowed on the apps at all. (about.fb.com) ### What actually changed? Meta already used AI to guess whether someone lied about their age. The update is that those checks are expanding across both Instagram and Facebook, and they now include “visual analysis” alongside the older signals like profile details, birthday messages from friends, and how people interact on the apps. Meta framed this as a broader age-assurance push, not a one-off test. (about.fb.com) ### What is the AI looking for? Basically, the system is trying to estimate an age range, not identify a specific person. Meta says it looks for general visual cues — things like height and bone structure — plus non-visual patterns that often give age away. That is why the company keeps saying this is “not facial recognition.” The distinction matters because facial recognition usually mea(about.fb.com)age estimation. (thehill.com) ### So is Meta scanning everyone’s face? Not in the “upload a passport-style scan for every login” sense. What Meta described is a risk-based system that flags accounts when its models think the stated age may be wrong. If the system is confident enough, Meta can move a user into age-appropriate settings. In some cases, the person may be asked to verify age with a video selfie or an ID. (about.fb.com) ### What happens if the AI thinks you are younger? For teens, the likely outcome is more restrictions. Meta’s Teen Accounts lock in tighter defaults around who can contact a teen, what content they can see, and how much of the app can nudge them to keep scrolling. For children under 13, the bigger risk is removal, because Meta’s rules do not allow them to hold accounts at all. (about.fb. ([about.fb.com)is Meta doing this now? Because the pressure has shifted from “say you care about teen safety” to “prove the age gate works.” Meta has been rolling out Teen Accounts since September 2024, expanded them to Facebook and Messenger in April 2025, and has spent the last year trying to catch users who signed up with adult birthdays. This week’s update is the next enforcement step — les(about.fb.com)ctually end up inside them. (about.fb.com) ### Why is the “not facial recognition” line doing so much work? Because the privacy fight is obvious. If a platform says it is analyzing faces, people hear surveillance. Meta is trying to narrow the claim: the company says it is not identifying who you are, just estimating roughly how old you look. But the catch is that, from a user’s point of view, the practical effect can still feel s(about.fb.com)extra friction, or removal. That distinction may satisfy some regulators, but probably not all critics. (thehill.com) ### Does this solve age verification? Not really. It solves one problem — obvious age lying inside Meta’s own apps. But it does not create a universal proof of age across the internet, and it still leaves room for false positives, appeals, and privacy complaints. That is why this story also points to a bigger industry argument: whether app stores or (thehill.com)e. That last part is an inference, but it follows from the direction of the policy fight and Meta’s own push toward automated enforcement. (9to5mac.com) ### Bottom line? Meta is moving from honor-system birthdays to automated age policing. That will probably put more real teens into safer default settings. But it also means social apps are getting more comfortable making sensitive decisions from behavioral and visual clues — and that shift is bigger than this one product update.