New AI Underage Detection Rolls Out on Instagram and Facebook

Meta has announced that it will use artificial intelligence to analyze photos and videos for physical indicators, such as height and bone structure, to help identify users who may be under 13 years old. The system aims to detect and remove accounts belonging to children who are not permitted on Facebook or Instagram.

The company emphasized that the technology does not use facial recognition. Instead, it examines general visual themes and cues combined with analysis of text, profile information, interactions, and other contextual signals like mentions of school or birthdays. Accounts flagged as potentially underage will be deactivated, requiring the user to complete an age verification process to restore access.

The visual analysis system is currently operating in select countries with plans for wider rollout. Meta intends to expand the technology to additional areas of its platforms, including Instagram Live and Facebook Groups.

This move is part of Meta’s ongoing efforts to address child safety concerns on its platforms. The announcement follows a recent New Mexico court ruling ordering Meta to pay $375 million in penalties for allegedly misleading users about platform safety and endangering children. Meta has faced multiple lawsuits related to child protection issues.

In a separate update, Meta is expanding its “Teen Accounts” feature on Instagram, which applies stricter safeguards for users identified as teenagers. These accounts limit direct messages to people the user already follows, hide potentially harmful comments, and set profiles to private by default. The expansion covers 27 additional countries in the EU and Brazil.