Meta Rolls Out AI to Detect Under‑13 Users Without Facial Recognition

Meta announced a new age‑verification system that will automatically examine content across Facebook and Instagram. The AI reviews images, video frames, captions, comments and other interactions to infer a user’s age. When the system estimates a user is under the legal minimum, the account is deactivated.

The head of North American Safety explained that the tool looks for clues such as birthday mentions, school grade references and visual indicators of stature. By aggregating these signals, the AI builds a profile that goes beyond a self‑declared age.

Crucially, Meta insists the technology does not employ facial recognition. Instead, it evaluates height and bone‑structure patterns to estimate how old a person appears. This distinction sidesteps privacy concerns tied to identifying faces.

The move follows increasing regulatory pressure to protect minors online. The United States and several states have introduced or enforced age‑verification requirements for social platforms. Meta’s approach aims to meet those mandates while avoiding the backlash associated with facial‑scan databases.

If the system works as described, it could reduce the number of under‑13 accounts that slip through manual checks. However, critics note that AI estimates can be error‑prone, potentially flagging older users or missing younger ones who conceal their age well.

The rollout will begin with a pilot on select accounts before expanding platform‑wide. Meta plans to refine the model using feedback and to publish transparency reports on removals.

What to watch: how accurately the AI distinguishes age, the volume of accounts it disables, and whether regulators deem the method sufficient to protect children online.