Meta is stepping up its efforts to keep under-13s off Facebook and Instagram. It’s now turning to AI-powered visual analysis, including cues like bone structure and height, to do it.
The company says it’s combining this with existing detection methods, such as scanning profiles, captions and interactions for hints about a user’s age. That includes things like mentions of school years or birthday posts, but the newer layer goes further. It analyses photos and videos to estimate whether someone might be underage.
Meta is keen to stress that this isn’t facial recognition. Instead, it says the system looks for “general themes and visual cues” to work out an approximate age range, rather than identifying a specific person. The idea is that combining visual signals with text and behavioural data improves its chances of spotting accounts that shouldn’t exist in the first place.
If Meta flags an account as potentially under 13, it will deactivate the account. Users must then provide proof of age to restore access; otherwise, Meta removes the account entirely.
This visual analysis system is currently being tested in select countries, with a wider rollout expected over time. At the same time, Meta is expanding tools aimed at teenagers. The system can automatically place users aged 13 to 15 into teen accounts, which include added protections and parental controls.
Those safeguards are now being extended more broadly. Instagram is rolling out this age-detection tech across Brazil and 27 EU countries. Meanwhile, Facebook is getting similar systems in the US first, followed by the EU and UK. WhatsApp has also introduced parent-managed accounts to allow younger users to access the app under supervision.
The push comes as Meta faces increasing regulatory pressure to better protect children online. The European Commission recently flagged concerns that the company may not be doing enough under the Digital Services Act, adding urgency to these updates.
