Meta expands AI age checks for teens

This is a meaningful escalation in platform safety automation: Meta is moving age enforcement from self-declared birthdays to probabilistic inference across text and images, which will improve scale but raise the false-positive and privacy debate immediately.

• –The system combines profile language, post context, and visual cues to infer age, then deactivates suspected under-13 accounts until proof of age is provided
• –Meta’s insistence that this is not facial recognition will not stop scrutiny, because “bone structure” and “height” are still sensitive inference signals
• –For product teams, this is a model for large-scale age gating: cheap, automated, and operationally effective, but hard to defend if it misclassifies adults or kids
• –The expansion to Facebook in the US, plus Instagram in the EU and Brazil, shows Meta treating age assurance as core platform infrastructure rather than a narrow policy tweak
• –The bigger industry signal is that app-level age verification pressure is rising, and Meta is using AI to argue for centralized age checks instead of per-app compliance