Australia's world-first social media ban for under-16s was meant to prove that age checks could protect children online. Instead, teenagers are already showing how quickly policy breaks when the technology is easy to dodge.
The ban took effect on December 10, 2025, forcing platforms including Instagram, TikTok, Snapchat, and X to stop new under-16 accounts and remove existing ones. On paper, the penalties look serious. Companies can face fines of up to $49.5 million for non-compliance, and more than 4.7 million accounts were reportedly removed in the first month. But the early evidence points to a very different reality: the rule is hard to enforce once teenagers, parents, and platforms all have incentives to keep the experience moving.
A survey by the Molly Rose Foundation of 1,050 Australian teenagers aged 12 to 15 found that more than 60% were still accessing at least one social platform after the ban began. TikTok, YouTube, and Instagram each retained more than half of their underage users, while two-thirds of respondents said platforms had taken little or no action against accounts created before the rules came in. That matters because this was not designed as a symbolic policy. Australia positioned it as a practical model for the rest of the world. If enforcement is this porous, other governments will have to think carefully before copying it.
Teenagers have turned the ban into a cat-and-mouse game. Evelyn, a 14-year-old from New South Wales, reportedly used her mother's face ID to keep using Snapchat and Instagram. Reddit threads have described mesh masks bought from Temu, makeup tricks, better lighting, and other small workarounds aimed at confusing facial scans. University of Melbourne tests found that $22 Halloween masks could bypass leading age estimation tools. Parents are part of the problem too. A May government poll found that one-third planned to help their children evade the ban, which tells you how far the policy is from real household behavior.
A Global Warning for Verification Builders
The technology stack is also under pressure. Platforms are using a mix of facial analysis, ID uploads, and behavioral inference, but Australian law does not allow them to make government IDs mandatory. That creates a gap big enough for determined users to walk through. The eSafety Commissioner is now investigating Meta, Google, TikTok, and Snap for suspected breaches, including failures during account setup and over-reliance on self-declared ages. Age Verification Providers Association executive director Iain Corby has argued that the issue is platform execution rather than the basic limits of verification technology, but that distinction may not comfort regulators if children are still getting through.
This is why the Australian experiment matters beyond social media. For identity startups, trust and safety teams, and age assurance providers, the lesson is blunt: a check that works in a lab is not the same as a check that survives real users. Biometric systems can be spoofed. Document checks create privacy and access concerns. Behavioral signals can be distorted by VPNs, shared devices, and family intervention. The next generation of age assurance will need stronger re-verification, better fraud detection, and privacy-preserving methods such as zero-knowledge proofs that can confirm age without handing platforms more personal data than they need.
Regulators elsewhere are watching closely. The UK's Online Safety Act, EU child safety rules, and a growing set of US state proposals all depend on the same promise: platforms can identify younger users accurately enough to enforce different rules for them. Australia is now testing that promise in public. Platforms that treat age-gating as a compliance checkbox will face fines, bad headlines, and pressure from parents who want clearer answers. Companies that build verification into the product itself, with repeat checks and privacy safeguards, will have a stronger case. The next race is not simply to ban children from platforms. It is to build verification that still works when users try to get around it.
Also read: Disneyland's facial recognition opt-out sets the standard for AI identity products • Blue Origin's New Glenn is grounded after its first real customer mission ended with a satellite in the wrong orbit • Chainalysis is adding AI agents to its blockchain forensics platform to counter criminals who are already using AI
