Meta Expands AI Age Checks After Children Outsmarted Existing Tools

Meta is widening automated age enforcement on its platforms

Meta is strengthening age-verification and age-estimation systems on Instagram and Facebook after repeated cases in which children reportedly bypassed existing checks with simple tricks, including the use of a fake mustache. According to the supplied source text, the company is deploying AI tools that look for age-related signals in posts, comments, biographies, descriptions, images, and videos to identify users under 13 and remove their accounts.

The shift reflects a broader industry problem: self-reported age is easy to manipulate, while online platforms face growing pressure to prove that their safeguards for minors work in practice rather than only on paper. Meta’s answer is to use a wider set of signals, combining textual clues with what it calls visual cues such as height and bone structure.

How the system works

Meta says the new tools analyze contextual indicators that may reveal a user’s age. These include references to school years or birthday celebrations in text, as well as automated analysis of shared imagery. The company is careful to say that the system is not face recognition and is not designed to identify specific people. Instead, it is meant to estimate whether an account is likely being used by someone younger than platform rules allow.

If Meta suspects an account is being run by a child under 13, the account will be suspended. The user must then revalidate their age through the company’s procedures to regain access. If that does not happen, the profile will be permanently deleted.

Why Meta is changing course

The reported trigger is straightforward: traditional age gates have proven too easy to evade. When a system depends heavily on what users type into a form, it only works as well as the honesty of the person filling it out. For children motivated to join adult-oriented or teen-oriented social spaces, the barrier can be trivial. The source text says hundreds of children have managed to get around restrictions, underlining the practical weakness of existing methods.

Meta’s new approach is part of what the article describes as an AI-based security strategy. Rather than trusting a single declared birthdate, the company is trying to infer age from behavior, content, and physical presentation. That could improve detection rates, but it also introduces more complicated questions about error, privacy, and appeal processes.

Teen accounts are also part of the plan

The policy is not limited to children under 13. Meta also says it will expand technology that identifies users aged 13 to 15 and automatically place them into teen accounts. These accounts come with content restrictions and parental controls enabled by default, which the company presents as a safer baseline for younger teens.

That is an important distinction. The company is not simply trying to remove prohibited underage users. It is also trying to sort age-eligible teens into a more restrictive product environment. In effect, Meta is using automated age inference both for exclusion and for product-tier assignment.

The tradeoffs ahead

The obvious benefit is stronger enforcement of platform rules that already exist. If the system works well, fewer children under 13 will remain on services that are not supposed to host them, and more younger teens will end up with stricter default protections.

But the cost of stronger automated enforcement is the risk of false positives. Any system that estimates age from text and images can make mistakes, especially around ambiguous content or people whose appearance does not fit the model’s expectations. Meta’s requirement that suspended users revalidate age creates a backstop, but it also shifts burden onto legitimate users who may suddenly have to prove who they are.

A larger signal for the industry

Meta began using age-verification technology for Instagram users in 2024 in the United States, Australia, Canada, and the United Kingdom, according to the source text. The latest expansion shows the company moving past verification at signup or on demand and toward continuous surveillance of account signals to estimate age over time.

That is a significant operational change. It suggests major platforms are increasingly willing to use AI not only to moderate content, but to classify the people producing it. If Meta’s system proves effective, similar methods could spread across the social media industry. If it proves error-prone, it may sharpen regulatory scrutiny around how platforms treat identity, minors, and biometric-adjacent inference.

Either way, the message is clear: age gates based on self-report are no longer enough for platforms facing legal, political, and public pressure to keep children out or place them into more constrained experiences.

This article is based on reporting by Wired. Read the original article.