Australia’s Teen Ban Has a Measurement Problem
Matthew Leo · Published August 13, 2026 · Business & Policy
Removing 756,000 suspected teen accounts sounds like enforcement at industrial scale. It does not prove that Australia’s social-media age restriction is keeping teenagers off social media.
Meta said on August 13 that it deactivated 462,000 Instagram accounts and 294,000 Facebook accounts believed to belong to Australians under 16 between the period just before the law took effect and June 2026. The figures are substantial, but they measure actions taken by one company. They do not tell the public how many teenagers stayed away, returned with new accounts, moved to other services or were removed by mistake.
That distinction matters beyond Australia. Canadian governments are also considering how age assurance, platform duties and youth protections should work. Copying the headline policy without copying a rigorous measurement system would repeat the weakest part of the Australian experiment.
Australia regulates the account, not the teenager
Australia’s social-media age restrictions have required covered platforms since December 10, 2025 to take reasonable steps to prevent Australians under 16 from creating or keeping accounts. The rule applies to services including Facebook, Instagram, Snapchat, Threads, TikTok, Twitch, X, YouTube, Kick and Reddit.
The obligation sits with platforms, not children or parents. A young person is not fined for opening an account, and the law does not require every Australian to upload government identification. Services can combine declared birth dates, account behaviour, age estimation and other signals.
Reuters reported Meta’s removal figures alongside a harder result: government figures and independent studies found that more than eight in ten under-16s were still using social media during the first three months of the restriction. The periods are not identical, so the numbers should not be treated as a direct failure rate. Together, however, they show why removals are only one part of the story.
A large number can conceal repeated failure
An account-removal total can rise for several reasons. A platform may be finding more underage users, the same users may be registering repeatedly, or the service may be removing older users whose age was estimated incorrectly. Only the first explanation clearly represents improved detection.
Canada should demand a common reporting table if it introduces comparable duties:
| Measure | What it reveals |
|---|---|
| Unique users affected | Prevents repeat accounts from inflating the total |
| Successful re-registration rate | Shows whether removal creates a lasting barrier |
| Appeal and reversal rate | Identifies adults and eligible teens caught by mistake |
| Method used | Separates self-declared age, inference and document checks |
| Data retained | Shows the privacy cost of proving age |
| Movement to other services | Reveals whether risk was reduced or merely displaced |
| Changes in exposure to harm | Tests the policy’s actual purpose |
Every covered platform should report the same definitions and periods. Otherwise, a company that aggressively counts failed sign-up attempts could appear more effective than one that prevents the attempt before an account exists.
Age assurance creates its own privacy risk
Platforms need some way to distinguish a 15-year-old from a 16-year-old. That can involve face-based age estimation, government documents, payment records, mobile-account information or behavioural inference. Each method creates different error and privacy risks.
Canada’s privacy regulator has already published guidance for privacy-protective age assurance. Its direction is useful: collect the minimum information, be clear about purpose and retention, secure the data, test for bias and provide a meaningful way to challenge a result.
A Canadian system should therefore measure more than access. It should disclose how often a service asks for sensitive evidence, how long that evidence survives, whether a third party receives it and how many decisions are overturned. A policy intended to protect children should not quietly build a durable identity database for everyone.
Canada should define success before choosing a ban
The case for protecting young people online does not depend on pretending every platform is harmless until perfect research arrives. Recommendation systems, direct messaging, public metrics and persistent notifications can expose children to risks that deserve direct regulation.
But account prohibition is only one tool. Governments can also require safer default settings, limit contact from unknown adults, restrict targeted advertising, give researchers access to platform data and enforce rapid responses to exploitation. Those measures can help even when a teenager reaches the service through a browser, a borrowed account or a smaller platform outside the law.
Australia is now testing one of the world’s most ambitious approaches. Canada should watch the results closely, including its mistakes. Meta’s 756,000 removals show that the machinery is active. The public still needs evidence that the machinery is producing safer lives rather than a larger counter.
Tags: social media, child safety, age assurance, Meta, Canada