News · Brain & Mental Health
Australia banned under-16s from social media. Most kept using it.
Three months after the world's first national minimum age law took effect, a BMJ study found more than 85% of under-16s were still on the restricted platforms.
Based on a peer-reviewed observational study in the BMJ
- Researchers surveyed Australian adolescents aged 12 to under 17 just before the Social Media Minimum Age Act took effect and again about three months later, publishing in the BMJ.
- Follow-up data were available from 408 of the 436 adolescents recruited at baseline.
- More than 85% of participants under 16 were still using platforms covered by the Act at follow-up.
- Between 54% and 68% were doing so through their own accounts rather than a workaround.
- Of those, 66% reported encountering age verification, most often self-declared age (24-39%) or uploading a selfie (13-27%).
- The study used a sharp regression discontinuity design, comparing adolescents just under 16 with those just over.
- This is a three-month snapshot of the earliest phase, before enforcement had time to mature.
- It is an observational study of 408 adolescents, not a nationally representative sample.
Australia did what several governments have been discussing: legislated a minimum age of 16 for social media accounts and put the burden of enforcement on the platforms. A study in the BMJ measured what happened in the first three months, and the answer is a lesson in the distance between a law and a behavior.
The researchers set out to examine the early effect of a world first national policy on adolescent social media use, and to describe adolescent use of social media platforms subject to the Act, and adolescents’ experience of age verification strategies, any efforts to circumvent them, and whether anything replaced the time.
Measuring a threshold
The design exploits the fact that the law applies to a birthday.
The team ran a community based study conducted across Australia among adolescents aged between 12 and under 17 years at the time of implementation of the age restrictions, surveying them immediately before the Act took effect and roughly three months after.
Then they used a sharp regression discontinuity design - comparing adolescents just below 16 with those just above. Someone aged 15 years and 11 months and someone aged 16 years and one month are alike in essentially every way except which side of the rule they sit on. Any difference between them is attributable to the law rather than to maturity.
Retention was good. Follow-up data were available from 408 of the 436 adolescents recruited at baseline.
What the law did
Not much, in three months.
More than 85% of participants aged under 16 years reported using social media platforms subject to the Act at follow-up.
And they were not mostly doing it through borrowed logins or a sibling’s phone. It was predominately via use of their own accounts, somewhere between 54% and 68% of them - the exact thing the legislation prohibits.
The verification problem
The data on age checks explain a good deal.
Of those under-16s using their own accounts, 66% reported exposure to platform age verification. So the systems existed and most users met them.
The issue is what the checks consisted of. It was most commonly self-declared age, met by 24% to 39% - a form where you type in a birth year. Less common was uploading of a picture, a selfie, at 13% to 27%.
A self-declared age check stops nobody who does not wish to be stopped. Facial age estimation is harder to defeat, and it was the minority experience.
What the study cannot tell you
Three months is the beginning of a policy, not its steady state. Platforms had to build verification systems, find existing under-age accounts and remove them, and respond to regulatory pressure that accumulates over time rather than landing on day one. The authors present this as early effects for that reason.
The sample is 408 adolescents in a community-based study, not a nationally representative survey, and everything rests on self-report. Adolescents reporting on whether they are breaking a new rule may under-report, which would make these figures conservative rather than inflated.
The study also measures use, not wellbeing. Whether the law improves adolescent mental health is the question the policy was sold on, and it is not the question this design answers.
Why the finding is useful anyway
Every country debating this faces the same practical question before the philosophical ones: can platform-side age checks actually work?
Australia’s experiment is the first real evidence. Early on, with the checks that were actually deployed, more than eight in 10 affected adolescents were still using the platforms and most had kept their own accounts.
That is not proof the approach cannot work. It is a fairly clear indication that writing the rule is the easy part.
People also ask
What does the Australian law actually require?
The Online Safety Amendment (Social Media Minimum Age Act 2024) set a national minimum age of 16 for holding accounts on designated social media platforms, and requires those platforms to take reasonable measures to prevent under-16s from having accounts. The obligation sits with the platforms rather than with children or parents, and reasonable measures is the phrase carrying most of the weight.
What is a regression discontinuity design?
A method for studying the effect of a rule that switches on at a threshold. Adolescents just under 16 and just over 16 are near-identical in every respect except which side of the law they fall on, so comparing those two groups isolates the effect of the rule itself rather than the effect of being older. It is one of the stronger designs available when a randomized trial is impossible.
Why did the law have so little effect at first?
The study documents what happened rather than diagnosing why, but the age verification data are suggestive. The most common check was self-declared age, which anyone can defeat by typing a different number. Selfie-based estimation is harder to bypass but was encountered by a minority. Existing accounts also had to be found and removed, which takes time.
Does this mean the law has failed?
Too early to say, and the authors frame this explicitly as early effects. Three months is a short window for a policy that requires platforms to build and refine verification systems, and compliance pressure on companies tends to build rather than arrive at once. What the study establishes is a baseline: at the start, most affected adolescents were still using the platforms.
Why does this matter outside Australia?
Because several countries are weighing similar legislation and Australia went first, making it the natural test case. The findings offer an early read on the practical question every such proposal faces, which is whether platform-side age checks can actually keep determined teenagers out. On this evidence, the early answer is mostly not.
References
- Barnes C, Hall A, Mantach S, et al. Assessing early effects of Australia's Social Media Minimum Age Act on adolescents' social media use: observational study. BMJ (2026).
- Centers for Disease Control and Prevention. Social Media and Youth Mental Health.
- National Institute of Mental Health. Child and Adolescent Mental Health.