A law can change account access before anyone knows whether it changed harm. Australia's first evaluation result measures the first step, while the larger promise remains open.
What the rule actually requires
From 10 December 2025, age-restricted social media platforms have had to take reasonable steps to prevent Australians under 16 from holding accounts. The obligation sits with platforms. It is often called a ban, but it is not a criminal penalty imposed on a child or parent.
That wording matters because reasonable steps are not a promise of perfect exclusion. Platforms choose combinations of age verification, estimation and inference. The regulator then assesses whether those steps comply with the law.
The first measured change
eSafety's first evaluation insights report that the proportion of under-16s holding a social media account fell from 52.4 per cent to 42.1 per cent. The 10.3 percentage point fall was statistically significant. That is evidence of a change in reported account access after commencement.
The result deserves to be reported without being inflated. A statistically significant change is not automatically caused only by the law, and account-holding is not the same outcome as exposure to harmful material, time spent online or wellbeing.
Access is the first link in a longer chain
The policy theory is that fewer under-16 accounts will reduce exposure to harmful platform features and content, which may improve safety and wellbeing. Each arrow in that chain needs evidence. A person can use some services without an account, move to another service, share an account or misstate an age.
Those possibilities do not prove the rule failed. They identify outcomes the evaluation must measure. The first number tells us the declared-account landscape changed. It does not yet show the size or direction of every downstream effect.
eSafety draws the same boundary
The regulator says the data were collected in March 2026, only three months after the obligation took effect, and that it was too early to meaningfully assess whether the intended outcomes had been achieved. A publication declaring the policy a safety success or failure now would go beyond its primary source.
Privacy and exclusion are part of the test
Age assurance can ask a service to infer or verify something about every user, not only children. Good implementation minimises data collection, protects the information used and provides a fair way to correct an error. A system that blocks adults incorrectly or stores sensitive identity data badly creates its own harm.
The opposite risk is weak assurance that leaves the policy easy to avoid. The regulator's compliance work has already focused on whether platforms provide effective checks and reporting pathways. Enforcement quality is therefore an outcome, not a footnote.
What the next evidence must show
Later evaluation waves should test whether the account reduction lasts, which services young people use instead, whether harmful exposure changes, whether wellbeing improves or worsens, how often legitimate users are blocked and what personal data the systems collect. Results should be separated by age and vulnerability where the sample supports it.
The first result says fewer under-16s reported accounts. It does not yet say the same children became safer.
The conclusion
Did the rule work? It produced an early, measurable movement in the intended access direction: reported account holding fell by 10.3 percentage points. That is more than no result.
Did it deliver the promised safety outcome? The available evidence cannot answer yet, and eSafety says so. Australia, Actually will review this article after the next substantive evaluation release rather than freeze an early verdict into a permanent headline.
- eSafety Commissioner: Social media age restrictions, Current obligation, platform responsibility and scope.
- eSafety Commissioner: Early insights from the evaluation project, 52.4 to 42.1 per cent account-holding result and the regulator's early-evidence limitation.
- eSafety Commissioner: Social media age restrictions evaluation, Intended and unintended outcomes the continuing evaluation will assess.
- eSafety Commissioner: Social media age restriction FAQs, Practical scope and responsibility clarifications.
Sources checked 2 August 2026. This is a developing-evidence article. The 10.3 percentage point decline is an early account-holding result, not a measured safety or wellbeing effect. Found a problem? See our correction process.