Meta will use AI to detect kids by analyzing height and bone structure

Dubai: Meta is turning to artificial intelligence to solve one of social media’s biggest and most controversial problems: figuring out who is actually old enough to use its platforms.

The company says it will begin using AI systems to analyse visual cues such as height, bone structure and facial proportions in photos and videos to identify users who may be underage on Facebook and Instagram. The move marks a major expansion of Meta’s age-verification efforts as governments worldwide intensify pressure on technology companies to better protect children online.

According to TechCrunch, the AI system will scan content for what Meta describes as “general themes and visual cues” that may indicate a user is under 13 — the minimum age required to use its platforms in many regions. Meta says the system is not facial recognition technology and does not identify specific individuals, but instead estimates age ranges using physical and contextual indicators.

The technology will also analyse captions, bios, comments and interactions alongside visual signals to determine whether an account likely belongs to a child. Accounts flagged as potentially underage may be suspended unless users can verify their age through official checks.

The rollout comes at a time of mounting scrutiny over children’s safety on social media.

Lawmakers in the United States, Europe and Australia have been pushing for stricter online age restrictions, while regulators increasingly argue that platforms relying on self-reported birthdays are failing to keep children out of age-inappropriate spaces. Reuters reported this week that Meta is expanding teen safety systems across Facebook and Instagram in Europe and the US, including broader AI-powered “Teen Account” protections.

Australia, meanwhile, is preparing tougher rules around social media access for minors, with Meta confirming its AI detection tools will play a role in enforcement efforts there.

The move also highlights how AI is becoming deeply embedded into moderation and identity systems across major platforms.

For years, social media companies relied largely on user reports and manual moderation. Now, increasingly sophisticated AI systems are being trained to infer age, behaviour and risk patterns automatically — a shift that could reshape how online identity verification works across the industry.

But the approach is likely to raise new questions around privacy, bias and accuracy.

Critics have already questioned whether physical appearance alone can reliably distinguish between a 12-year-old and a teenager just above the minimum age threshold.

For Meta, however, the message is increasingly clear: the era of simply asking users to enter their birthday may be coming to an end.