Meta has a new way of catching kids who lie about their age online, and it goes well beyond checking what they type.
The company announced it is now using AI visual analysis to scan photos and videos on Instagram and Facebook for physical indicators of age, including height and bone structure. Meta’s goal is to find and remove accounts belonging to users under 13 who may have signed up with a false birthday.
How does the visual analysis actually work?
Meta has been careful to clarify that this is not facial recognition. The AI doesn’t identify who someone is. Instead, it scans for general visual cues that suggest someone is young, like their physical proportions, to estimate a broad age range.
This visual scan works alongside existing text-based detection, which looks for contextual clues like birthday mentions, references to school grades, and information in bios, posts, captions, and comments. Meta also plans to expand this text analysis to Instagram Reels, Instagram Live, and Facebook Groups.
If an account is flagged as potentially underage, it gets deactivated. The user then needs to verify their age to get it back, or the account gets permanently deleted. The visual analysis is currently live in select countries, with a broader rollout planned.
What else is Meta doing for teen safety?
Meta is also expanding its Teen Accounts system, which automatically places users it suspects are between 13 and 15 into a stricter account experience. That means private accounts by default, DMs limited to people they already know, and hidden harmful comments.
This expansion now covers Instagram in Brazil and 27 EU countries, following earlier content restrictions modeled on film ratings. Facebook in the US is getting it for the first time too, with the UK and EU following in June. Meta has also given parents visibility into their kids’ AI chats as part of the same broader push.
The moves come as Meta faces mounting legal and regulatory pressure over child safety, including a $375 million penalty in New Mexico and a European Commission investigation into whether its platforms are doing enough to keep children off them.
