Instagram will use AI to catch teens lying about their age

Lying about your age on the internet used to be the easiest trick in the book. A simple click on "I am over 18," and you were in.

On Instagram, that era is about to come to a grinding halt.

Meta is rolling out a proprietary AI tool called the "adult classifier" early next year. Its mission is simple: sniff out teenagers lying about their birthdates and automatically herd them into restrictive Teen Accounts.

The AI reading your birthday wishes

According to Allison Hartnett, Meta’s director of product management for youth and social impact, the software splits users into two buckets: over or under 18.

To make the call, the AI digs through profile details, follower lists, engagement patterns, and even casual "happy birthday" posts from friends. If your bio claims you are 21 but your friend's story says "Happy 15th," the algorithm will flag you.

Meta is also tracking device IDs and flagging accounts created under existing email addresses with shifted birthdays. The old workaround of having a friend vouch for your age has already been axed.

Once flagged, teens under 18 are shoved into restrictive settings that limit DMs and hide sensitive content. Those aged 16 and 17 can change settings back themselves, but users under 16 will need explicit parental consent.

For adults wrongly flagged as minors, proving your real age will require uploading a formal ID like a driver’s license or submitting a video selfie to third-party estimator Yoti.

Backed into a regulatory corner

Meta’s sudden urgency isn't happening in a vacuum. A study by the UK’s telecommunications regulator found that a full third of minors on social media falsely claim to be adults.

Meanwhile, Meta is facing lawsuits from dozens of state attorneys general accusing it of hooking kids and fueling a teen mental health crisis. Ever since whistleblower Frances Haugen leaked internal research in 2021 showing Instagram's toxic effect on teenage girls, the platform has been under intense fire.

Meta hasn't revealed the classifier's accuracy rate yet. An official appeal process is still in the works, though miscategorized adults can currently toggle restrictions off without parental permission while the system is fine-tuned.

Why it matters

We think this is the beginning of the end for the internet's honor system.

For decades, tech companies relied on passive self-reporting to maintain legal plausible deniability. Meta deploying AI to actively police user age will force the rest of the industry, from TikTok to Snap, to abandon self-attestation and follow suit.

For users, the trade-off is clear: keeping teens safe will come at the expense of privacy. If the algorithm misjudges your browsing habits or friend group, you will be forced to hand over government IDs or biometric facial scans to regain normal access.

Meta's tactic is clever because it leverages existing social graph data without requiring new hardware. But having an advertising giant routinely scan personal interactions and birthday messages under the banner of child safety sets a privacy precedent that looks uncomfortably invasive.