Do social media age limits work? What Australia actually shows

Short answer: an age limit works as prevention for the children who have not gone in yet, and barely at all for those already inside. Restrictions on how platforms treat minors work better on the second group, but do not solve the problem on their own.

I am not going to tell you which side of this debate is right. I think both camps are arguing in good faith, and that they disagree about something real. What I can contribute is two things: what the numbers actually show, and what this looks like when you read it as communication rather than as politics.

The second is my job. I have spent years looking at what a message does, as distinct from what it says.

What Australia actually shows

Australia introduced the world's first national 16-year limit on 10 December 2025. Norway is now debating what is in practice the same model. So this is the closest thing we have to an experiment with ten months of results in.

The numbers are mixed, and they point in both directions. Here is the ledger, as plainly as I can set it out.

Why "it failed" is too imprecise

It is tempting to read the above as a crushing verdict. I do not think that is an honest reading, and I think the imprecise version loses arguments it ought to win.

The law failed at the thing it was named after. It was meant to keep under-16s off the platforms, and it did not.

It also did three other things that are real. It moved responsibility from children and parents to the companies, which is a shift in principle regardless of compliance. It gave a generation of parents a language and a backstop they did not have: "it isn't allowed" is a different starting point from "I don't want you to." And it did lower time spent in one age group.

So the precise sentence is this: Australia succeeded at changing the norm and failed at changing access.

That distinction is not pedantry. It determines what one should reasonably expect from the Norwegian law.

Norway is going the same way, and the government knows it

The government aims to put its bill before the Storting before the end of 2026. The limit attaches to 1 January of the year the child turns 16, so whole year groups gain access at once. Enforcement is anchored in the EU's Digital Services Act, and the duty falls on platforms through age verification requirements. The EU gave positive feedback on the proposal in August.

Here is what I think is the most overlooked sentence in the entire Norwegian debate, and it appears in the government's own bill:

The government states that an age limit on social media will not be sufficient to prevent children from being on social media.

That is not a concession extracted by an opponent. That is the proposer describing the instrument's reach correctly.

And it is worth noting that those critical of the implementation are largely not opposed to the goal. The Norwegian Data Protection Authority supports the intention, and director Line Coll has pointed out that a statutory age limit can in itself be norm-forming and carry an important signal effect toward children, parents and providers.

That is precisely what Australia achieved. The authority's objections concern something else: that the age verification requirement constitutes an intrusion into the privacy of everyone who uses social media, not only children, and that the consequences have not been adequately assessed. Section head Tobias Judin has put it precisely: age verification can always be circumvented by those who really want to.

The authority's own proposal deserves more attention than it has had: introduce the age limit first, observe its effect, and only then consider the more intrusive verification requirement.

That is a proposal which takes the signal effect seriously without paying the whole privacy bill up front. I have not seen a good counter-argument to it.

The second track: the letter to Meta

Alongside the law, something else is happening, and this is where the discussion gets interesting.

Following Meta's settlement with US states, Minister for Children and Families Lene Vågslid sent the company a letter on 10 September 2026 expecting the same measures for Norwegian children. Those measures include a maximum two hours of combined daily use of Facebook and Instagram for minors, restrictions on notifications at night and during school hours, the option of an algorithm-free content feed, fewer notifications and hidden like counts.

Prime Minister Jonas Gahr Støre has justified it on the grounds that the state has a responsibility to protect children from harmful screen use. Vågslid has emphasised that the measures can also protect children not covered by the planned age limit.

That is a good point, and it is the strongest argument for this track: restrictions reach the children who are already inside. An age limit does not.

Now to what I notice as a communications professional, and this is an observation about the instrument rather than about the politics: according to Verdinytt, the government had not assessed the individual measures separately before sending the letter, the letter contains no deadline, it does not ask for a timeline, and Meta had not replied when the matter was reported in mid-September.

A letter with no deadline and no requested timeline is not regulation. It is an expectation expressed in public.

That is not necessarily wrong. A publicly expressed expectation is a legitimate and sometimes effective instrument, particularly against a company in the middle of a reputational problem. But it is a different instrument from a law, with a different mechanism and an entirely different time horizon, and it should be read as such.

The only real mistake available here would be confusing the two.

The China comparison, done properly

Several people have pointed out that the list of measures resembles how Douyin, the Chinese version of TikTok, is regulated. That comparison is worth taking seriously, and it is worth doing accurately, because it is used loosely.

Here is how Douyin's youth mode works: users under 14 are placed in it automatically, use is capped at 40 minutes a day, the app is unavailable between 10pm and 6am, and the content inside the mode is curated toward science, history and museum material. The mode became mandatory in 2021, after China's Minor Protection Law was revised to require services to provide functions for time management, content restriction and spending limits for minors.

The similarity in mechanics is real. Time cap, night curfew, control over what the feed serves. Those are the same three levers.

But two things should be said, and they pull in opposite directions.

The first is that the comparison is often overstated. TikTok outside China already applies a default 60-minute daily limit for users under 18. Western platforms have had equivalent functions for some time. The major difference between China and the West has not primarily been the functions, but whether anyone enforces them.

The second is that there is nevertheless a difference that matters, and it is not the time limit. It is the content curation. Douyin's youth mode has banned categories including superstition and the "subversion of traditional moral values." There, a state has taken a position on what is wholesome for a child to see.

The Norwegian requests concern mechanisms (time, notifications, the option to escape the algorithm) and not which content is good for children. That is a real and important difference, and it should be acknowledged rather than waved away.

At the same time: the moment a state begins asking a platform to shape what minors see, the question of who decides what belongs there is in play. Not settled. In play.

I am not suggesting Norway is heading toward Chinese conditions. I am suggesting that the distance between "fewer notifications at night" and "curated content for minors" is shorter than it feels, and that this is a conversation worth having with open eyes rather than in retrospect.

It is also worth listening to those it concerns. The Norwegian Children and Youth Council has put its objection briefly: do not wrap us in cotton wool and call it protection.

What the discussion is really about

When I put the two tracks side by side, I see one shared movement, and it matters more than which model one prefers.

Both move responsibility away from the adults around the child. The age limit moves it to the platform. The restriction model moves it to the state, which moves it onward to the platform.

That is progress in itself. It was unreasonable for the responsibility to sit with a twelve-year-old and a parent who does not know the app. Norway's media regulator has also been clear that regulation must target functions and design capable of causing harm, rather than what a service is called — a considerably better criterion than a list.

But there is a limit to what that movement can deliver, and Australia has now shown where it lies.

A child who is on a platform that is not on the list, at 23:40, with a message from someone she does not know, is helped by neither. She is helped by there being a trusted adult she tells.

That is not an argument against regulation. I am in favour of this law arriving, and I think the restriction track is sensible. It is an argument about sequence and realism: regulation changes the conditions, and a trusted adult changes the outcome.

The government has, as noted, written exactly this in its own bill. I only wish that sentence got as much room in the debate as the age limit does.

We have written more about why a trusted adult is the one thing that follows a child from app to app

The short answer

An age limit is prevention. It helps the children who have not gone in yet, and Australia shows it works better as a norm than as a barrier.

Restrictions are harm reduction. They reach the children already inside, which is the group the age limit does not.

Neither is guidance, and guidance is what determines what happens when a child meets something difficult.

A law can decide who gets into the room. It cannot be in the room.

For those of you who work with this

For municipalities, child protection services and the youth sector, this debate has a practical consequence that is easy to miss.

However the law lands, it takes effect in 2027 at the earliest, and it will have a bedding-in period. The Australian experience suggests compliance will not be complete for some time. The children you meet over the next two years will largely be on the same platforms as today.

That means planning which assumes the law solves the problem is planning for the wrong year.

The useful question in a staff meeting is not which apps will become legal. It is whether you have a shared language for what happens inside any app, and whether each individual child has a trusted adult in the digital room whose name they know.

Norway's NOU 2024:20 places digital care with the adults around the child. It does not specify which apps, and it is not waiting for a law.

Frequently asked questions