Beyond birthdates: TikTok’s new age-detection push 📱

Cliquez ici pour lire en français

Cyberbullying, inappropriate content, screen addiction, mental health harm: the risks tied to early social media use are now well documented. Yet despite the age restrictions most platforms already claim to enforce, the actual controls have largely failed to hold up.

Until now, most social networks simply asked users for their date of birth at sign-up — a method as simple as it is easy to game. Against that backdrop, the tech industry is increasingly turning to artificial intelligence and behavioral analysis to better identify underage users.

A problem climbing the policy agenda 🌐

Regulators worldwide are stepping up scrutiny. In France, social media access for under-15s still requires parental consent. A bill that would have banned it outright was passed by Parliament in July 2026, but it was struck down by the Constitutional Council before taking effect — the matter remains under debate. Australia went further still: since December 2025, the country has enforced a full ban on social media for anyone under 16 — a world first.

These moves reflect a growing question: how do you meaningfully protect minors on platforms built to maximize engagement? Mounting debate over the psychological effects of social media has pushed the search for technical solutions that can actually enforce age limits, rather than just declare them.

TikTok rolls out age-estimation tech 🤖

That’s the backdrop for TikTok’s announcement that it will roll out an AI-based system across Europe. The goal: flag accounts likely to belong to children under 13, even when the birthdate on file says otherwise.

Unlike traditional checks, this system doesn’t rely on self-reported data alone. The algorithm reportedly weighs a mix of signals — account activity, browsing behavior, content interactions, and other usage patterns — to estimate a user’s likely age.

The shift illustrates a broader trend in content moderation: platforms are moving from a declarative model (trust what users say) to a predictive one, built on data analysis and machine learning.

Innovation meets data protection concerns 🔐

When an account is flagged as potentially belonging to an underage user, the case reportedly goes to specialist human moderators for review. Users keep the right to appeal and prove their actual age.

Several verification methods are on the table: facial age estimation through third-party providers, ID document checks, or credit card verification. But the approach raises its own set of questions. Data privacy, algorithmic reliability, and the risk of false positives remain top concerns among digital rights experts.

With the EU’s Digital Services Act tightening platform obligations, this new generation of tools could mark a turning point for child safety online. Whether AI can actually balance effectiveness, transparency, and privacy protection, though, remains an open question.

Do you think AI is the most effective way to keep minors off social media before the legal age? Let us know in the comments.