A father checks his fourteen-year-old daughter’s phone, convinced the new law will keep her safe. The official account has vanished, yes. But on the screen appears another: a different name, a fabricated date of birth. The prohibition exists. The access, too.
That scene condenses the knot at the heart of one of the most heated debates about childhood and technology. For years it has been assumed that simply restricting entry to social networks by age would safeguard the mental health of minors. The first robust data point points to something less comforting: bans act as an initial barrier, but they do not solve the problem by themselves.
Australia as a Reality Check
The first serious thermometer arrived on June 24, 2026. On that day, BMJ published a study led by Barnes and colleagues that followed adolescents aged roughly 12 to nearly 17. The team surveyed them just before the restriction took effect and again about three months later. They started with 436 participants; 408 completed the follow-up, about 94%. The data most cited since then reveal a discrepancy between what the law requires and what teens actually do: more than 86 % of those under 16 reported having used, in the preceding week, at least one platform covered by the rule. Nor did there appear to be strong evidence that daily usage time had fallen substantially.
That result makes sense when looking at the architecture of the law. The most ambitious regulatory experiment in the world is under way in Australia. Since 10 December 2025, the Online Safety Act 2021, in its Part 4A, Section 63D, requires social platforms subject to the regulation to adopt “reasonable measures” to prevent minors under 16 years from creating or retaining accounts. The eSafety Commissioner, Australia’s digital-safety authority, insists on one point: responsibility lies with the companies, not with the children or their families.
To ensure the message sticks, the law threatens civil penalties for systemic non-compliance. The official 2025 estimate set the maximum fine for a company at AUD 49.5 million, equivalent to 150,000 penalty units of AUD 330 each. In 2026, after raising that unit to AUD 364, the figure climbed to AUD 54.6 million. These sums are designed to touch the bottom line of companies accustomed to operating at a global scale.
The figure of 86 % helps explain why the regulator lowers expectations: the rule does not erase access to the internet. Its effectiveness hinges on age verification, on platforms detecting minors, and on their ability to curb fake accounts or evasion of controls. That is why the eSafety Commissioner will assess implementation over at least two years. The 2026 data are early compliance results, not a verdict on what the law does to child well-being.
What Early Data Reveal: The Law Exists, Access Persists
Does this mean the policy has failed? Not necessarily. It is a snapshot taken only three months after the law’s application, in a relatively small sample. The regression discontinuity design compares with precision those who fall just below and just above the 16-year threshold, and that is its value. What it clearly shows is a gap between the legal prohibition and observable behavior. What it cannot yet measure are medium-term changes in harassment, exposure to harmful content, sleep, or mental health.
The Displacement Problem: Fewer Accounts, Not Less Risk
This is where the question most likely to worry families arises: it’s not about how many accounts are deleted, but what happens to the minors who remain online. The eSafety Commissioner has flagged a tricky statistical nuance. The figures for removed accounts do not equal the number of protected children, since one person may hold multiple profiles and some accounts may be inactive.
Consider a teen who migrates to a fake account, to a service not covered by the rule, or to a private space. That minor becomes less visible to parents, teachers, and content moderators. The prohibition may curb certain contacts, but it shifts part of the risk into corners where nobody is watching. Removing the signal does not automatically extinguish the danger.
Mental Health: A Real Connection, But Misunderstood
A large share of the push toward prohibitions stems from concerns about adolescent mental health. The evidence calls for caution. The systematic review and meta-analysis by Fassi and colleagues, published in JAMA Pediatrics in 2024 and with a data correction in 2026, gathered 143 studies, nearly 1,095,000 adolescents, and 886 effect sizes.
The finding? There is a positive association between social media use and internalizing symptoms such as anxiety or depression, but of modest magnitude: r = 0.14. It is a small correlation that, moreover, does not establish causality. The study blends cross-sectional and longitudinal studies and uses very different measures of use—from total time to differences between active and passive participation. This overall number does not describe a single way of “using networks” nor prove that screens cause depression in every teenager.
The U.S. Surgeon General named it in a 2023 report as a complex and potentially bidirectional relationship: platforms can influence well-being, but preexisting mental health issues also shape how and how much a teen uses social networks. Its conclusion was prudent and eloquent: “We cannot conclude that social media are safe enough for children and adolescents.” It did not claim they are uniformly harmful, only that safeguards are lacking.
What Prohibitions Leave Out: The Benefits
The other side of the story should not be ignored. For many adolescents, social networks play protective roles: sustaining friendships, offering peer support, providing a channel for self-expression, and granting access to information that is hard to obtain otherwise. That value is especially high for isolated youths, for those with disabilities, for minorities, or for those living far from their usual support network.
A blanket ban cuts access blindly, and with it, those advantages. And another undercurrent becomes apparent: age is an administrative variable, not a reliable meter of maturity or vulnerability. A fifteen-year-old may thrive in a safe digital community while another sixteen-year-old may suffer intense harassment or encounter dangerous content. The threshold blocks by age, not by person.
A Suite of Measures More Ambitious Than a Simple Ban
If the prohibition is merely a barrier at the outset rather than a fortress, what should accompany it? Public health and digital-safety authorities point to measures that attack the concrete mechanisms of harm, not the mere fact of having an account.
The list is long: default-private accounts for minors; restricting messages from strangers and targeted advertising; controls on endless scrolling, notifications, and recommendation systems designed to keep you engaged longer. Add to that independent audits of algorithms and rapid removal of material linked to exploitation, self-harm, or eating disorders. No single approach has yet shown definitive proof of being the most effective across all contexts. They share, however, a common logic: intervene in the product’s design, not only at the gate.
The rest of the framework is human. Digital literacy for minors, families, and teachers. Parental controls that accompany without spying. Accessible mental health services. All of this sketches a holistic approach that fits the problem, which is multifactorial. And there is a tension seldom named: the stricter the age verification, the more personal data are collected. Regulation will have to prove that it protects children without building a surveillance-heavy system.
The answer to the opening question is not a clear yes or no. Age-based restrictions can help some minors and delay entry for younger ones, though no data show that a blanket prohibition, by itself, improves child mental health. For families, this translates into a practical consequence: trusting that a daughter is safe because her official account vanished means trusting the wrong lock. Age may be the first lock on the door; true safety, however, must be built into the design of the platforms on the other side.
