Australia banned under-16s from social media. Did it work?

On 10 December 2025, Australia switched off social media for everyone under 16. Malaysia followed on 1 June 2026. In his National Day Rally speech on 23 August 2026, Prime Minister Lawrence Wong said that on ages, “self-declaration is just not good enough”. Three governments in this region, pointing the same way, inside nine months.

So the question for a parent here is not whether the gate is coming. It is what happens to a child standing on the other side of it. Australia is the only country in the region that has run the experiment long enough to measure, and the first measurements are now public.

The first honest scorecard

Australia’s online safety regulator, eSafety, is running a two-year evaluation of the ban, with Stanford’s Social Media Lab as lead academic partner. The first report landed on 31 July 2026, covering 803 children aged 10 to 15 and their parents, surveyed at baseline in December 2025 and again in March and April 2026.

The share of children using an age-restricted platform fell from 85.9% before the ban to 81.5% three months after it. Account ownership dropped further, from 52.4% to 42.1%. The report’s own summary is not triumphant: reductions in children’s use and account ownership, it says, “have been limited in scale”.

The reason is not mysterious. Just over half of the children who still had accounts, 50.2%, said the platform had never once asked them to verify their age. Australia’s assistant minister Andrew Leigh was blunt about it in early August: “We never expected that this would have 100 percent compliance.”

Which is not to say nothing happened. Meta reported on 12 August that as of 30 June it had removed access to more than 750,000 Australian Facebook and Instagram accounts it assessed as belonging to under-16s, over 500,000 of those before the law even took effect. It says it finds them by using AI to read profiles for “contextual clues that an account may belong to someone under 16, such as birthday celebrations or mentions of school grades”. Hundreds of thousands of accounts really did close. The children behind them mostly did not go anywhere.

[[photo:1]]

The finding nobody was hoping for

Buried in the same evaluation is a number that deserves more attention than the headline one.

The proportion of parents who did not know their child had used social media recently rose from 23.3% to 33.3%, with higher rates among parents of girls and of children aged 10 to 12. Over the same period, daily-or-more use of messaging platforms climbed from 40.5% to 52.3%, with more time also going to games and to apps the ban does not cover.

Read those two together and you get the shape of what the ban did in its first quarter. It did not take children off the internet. It moved a slice of their day into rooms with fewer adults in them, and it left a third of parents less certain about where their child actually is.

Some care is due here. This is one evaluation at the three-month mark, the first of a series running into 2027, and the researchers say plainly that it is too early to call the policy a success or a failure. Habits take longer than a quarter to settle. But it is real evidence, gathered by the regulator with every reason to want good news, and it points one way rather than two.

Two designs, one region

Malaysia took a different route to the same destination. Since 1 June 2026, under the Child Protection Code and Risk Mitigation Code issued under the Online Safety Act 2025, licensed platforms cannot let an under-16 open an account, and they must check age against government records: MyKad, a passport, or MyDigital ID. Penalties run to RM10 million. Australia set the age and left the method to the platforms. Malaysia named the document.

That difference matters, because the Australian data suggests the weak link was never the ambition of the law. It was whether anybody ran a check at all.

What Singapore looks to be building instead

Singapore has not announced a ban. What it has announced is closer to a redesign.

Wong said in August that platforms “will have to put in place robust and reliable checks so that the minimum age is properly set and enforced”, and that where safeguards stay inadequate the government would consider raising the minimum age beyond 13 for that platform. He cited a figure worth sitting with: one in six young people aged 10 to 24 here shows signs of problematic social media use.

Since then the detail has firmed up. On 31 August, Minister for Digital Development and Information Josephine Teo set out safeguards under consideration for teenagers, among them daily time limits and curbs on infinite scroll, autoplay and messages from strangers. “The very nature of infinite scroll is that you don’t have a natural stop point,” she said. “You keep going on and on.” Legislation is expected to be tabled early in 2027. Back in May, MDDI had already told Parliament it would extend age assurance requirements to designated social media services and was studying features such as direct messaging and autoplay.

The interesting part is where the aim is pointed. An age gate regulates the child. A rule about autoplay and infinite scroll regulates the product. Given what the Australian numbers say about gates on their own, that is not a small distinction.

[[photo:2]]

The number to act on at home

Of everything in that evaluation, the one to sit with is 23.3 rising to 33.3. The policy did not put more parents in the picture. In its first three months it put fewer.

No bill fixes that, and no bill needs to. Ask about the messaging apps and the games, not only the famous platforms, because that is where the traffic went. Assume that when one door closes your child finds another, and treat that as ordinary rather than as a betrayal. Then protect the thing no regulator can legislate: being the person your child tells when something online goes wrong, without the telling costing them their phone.