Meta Deploys AI to Detect Underage Users via Height and Bone Structure Analysis
Summary
- • Meta launches AI scanning photos for height and bone structure to identify users under 13
- • Explicitly not facial recognition — system reads general physical proportions to estimate age
- • Flagged accounts are deactivated immediately; age verification required to avoid deletion
- • Teen Accounts expanding to 27 EU countries, Brazil, and U.S. Facebook for the first time
Details
AI analyzes height and bone structure in photos/videos to estimate if user is under 13
Meta explicitly states this is not facial recognition — the system reads general visual proportions, not individual identity. Combined with text and behavioral signals (birthday posts, school grade mentions) for higher detection accuracy.
System live in select countries now; expanding to Instagram Live and Facebook Groups
A broader global rollout is in progress. Extension to live video and group content significantly widens the system's detection scope beyond static posts and profiles.
Flagged accounts deactivated immediately; users must pass age verification to avoid permanent deletion
The enforcement flow raises open questions about false positives and burdens placed on legitimate young adult users misidentified by the physical-cues model.
Teen Accounts expanding to 27 EU countries and Brazil; coming to U.S. Facebook for the first time
Teen Accounts enforce stricter defaults: DMs limited to followed accounts, harmful comment filtering, private-by-default settings. U.K. and EU Facebook rollout follows in June 2026.
New Mexico jury ordered Meta to pay $375M in civil penalties for child safety failures
The penalty came weeks before this announcement, reflecting ongoing legal and regulatory pressure on Meta over its handling of minors. The timing of these safety announcements is notable.
Multi-signal approach combines visual cues, text analysis, and behavioral pattern analysis
Combining vision, NLP, and behavioral modalities is intended to reduce errors and increase coverage. This is an example of multi-modal AI deployed at scale for consequential moderation enforcement.
New Tech = novel AI capability, Product Launch = feature rollout, Policy = platform enforcement, Legal = litigation/regulatory action, Tech Info = technical implementation detail
What This Means
Meta is moving toward continuous, automated biometric-adjacent scanning of user-generated content to enforce age restrictions — a significant escalation in platform-level AI surveillance that sets a precedent for how social networks handle child safety compliance globally. For AI practitioners, the system illustrates how multi-modal models (vision + NLP + behavioral signals) are being deployed at scale for consequential, real-world moderation decisions with direct account-level enforcement. The $375M penalty and regulatory expansion into the EU suggest this is as much a compliance-driven initiative as a product one, and the false-positive risk for legitimate users flagged by the physical-cues model remains an open and unaddressed question.
Sentiment
Mostly concerned about privacy and surveillance, with limited support for child safety measures
“Age verification has reached the 'let the model inspect your skeleton through a selfie' phase. Very normal internet governance.”
“Meta's AI will analyze users' bone structure and height for age verification... Accounts flagged as minors are deactivated. The user is then forced through Meta's age verification process.”
“Their AI will scan posts, comments, bios... photos & videos for visual cues like height and bone structure... If flagged, your account gets deactivated until you submit ID... real ask is government mandate at app store level.”
notably framing it as lobbying for broader government-mandated age verification
“Child safety measure… or straight-up dystopian Big Tech body surveillance on millions of users? This crosses a dangerous line.”
Split
~80/20 concerned/privacy risks vs child safety benefits.