Under-16 social media bans: why they barely worked

On 10 December 2025 Australia did something no country had done before: it locked every person under 16 out of social media. Millions of teenagers went to bed with Instagram and TikTok and woke up without them.

Three months later the national regulator (the government body that enforces the law) surveyed 803 families. 81.5% of children aged 10 to 15 were still using social media. Before the ban the figure was 85.9%.

Four percentage points. That is the entire visible result of the largest disconnection in internet history.

It reads like a failure. But there is an explanation, and it is more interesting than the ban itself: a very similar experiment had already run in English schools, and it ended exactly the same way. For exactly the same reason.

What actually happened in Australia

Accounts disappeared and usage barely moved. That is the paradox of the Australian ban.

From the platform side the numbers look serious. By mid-January companies had removed, deactivated or restricted 4.7 million accounts. Meta reported that by 30 June 2026 it had taken down more than 750,000 Facebook and Instagram accounts in Australia, half a million of them before the law even came into force. Fines reach 33 million dollars per platform, so the companies took it seriously: they hunted suspicious profiles with AI, made it easier for parents to report accounts, and raised their app store age rating to 16+.

From the children's side it looks different. According to Qustodio (a parental control company that sees anonymised app usage data), by August 2026 TikTok was still being opened by 25.9% of Australians aged 13 to 15. That is barely one percentage point below the pre-ban level. Among 10 to 12 year olds the gap was 0.06 percentage points, which is no gap at all. Instagram and Snapchat did not fall in either group, they grew. WhatsApp, which the ban does not cover, was installed more often.

Millions of deleted accounts, four points of difference. How do those two facts fit together?

Why teenagers never left

Because more than half of them were never asked how old they were.

The regulator blamed the weak result on one thing: platforms failed to implement age assurance properly (age assurance means checking that a user really is the age they claim). The survey shows how that played out:

• more than 50% of children said their age was never checked at all;
• 37% simply typed in that they were 16 or older;
• 18% went through automatic facial age estimation and the system decided they were adults.

In most cases a child needed no VPN (software that makes it look like you are in another country) and no borrowed ID. A different number in the year-of-birth field was enough.

The second half of the answer is migration. Some teenagers moved to places the ban does not reach: Reddit, Pinterest, messaging apps, smaller platforms such as Lemon8. Others stopped logging in altogether, because you can scroll a feed in a browser without an account. Parent surveys found that close to 70% of previously active children still had accounts on Facebook, Instagram, Snapchat or TikTok.

Meanwhile complaints about bullying and image-based abuse among under-16s showed no clear decline.

So suppose age checks worked perfectly. Would children be better off? That question already has an answer, and it came not from Australia but from English schools.

English schools ran this experiment first

When you take the phone away for part of the day, screen time does not shrink. It moves to the evening.

In 2025 The Lancet Regional Health - Europe published the SMART Schools study. Researchers compared 30 English secondary schools: 20 banned phones during the school day, 10 allowed them. They surveyed 1,227 students.

Inside the school day the difference was clear. In strict schools teenagers spent 0.67 hours less on their phones and 0.54 hours less on social media. The rule worked.

What happens next is the part worth remembering. Outside school there was no difference at all. Total phone time and social media time on weekdays and weekends matched across both groups. Students from strict schools made up the difference in the evening. This is called compensation: a restriction does not remove a behaviour, it shifts it in time.

And on wellbeing the two groups were indistinguishable. No difference in the wellbeing scale, anxiety, depression, sleep, physical activity or grades. The authors wrote it plainly: their data do not support school phone bans in their current form, and such policies need further development.

The study has an honest limit that the authors state themselves: it is a snapshot in time, not a multi-year follow-up of the same children. That design shows association, not cause. Some phone-time data came from what teenagers reported about themselves. And fewer than 11% of invited schools agreed to take part, which risks a skewed sample.

Still, the match is too precise to ignore. Two different bans, two countries, one result: the time does not vanish, it relocates. Which raises the question most bills skip. Is time the problem at all?

The harm tracks how teenagers use social media, not how long

What lines up most closely with a teenager feeling bad is not hours on social media but a compulsive pattern of use.

In 2026 the journal European Psychiatry published a meta-analysis (a study that pools dozens of other studies and calculates a combined result). It covered 59 studies and 547,697 participants with a mean age of 16.6. The researchers separated two things that usually get lumped together.

• Problematic use (being unable to stop, lying about screen time, fighting with family over it, turning noticeably low when access is removed) links to psychological distress at 0.39.
• Time spent on social media links to distress at 0.15.

Those numbers are correlation coefficients. Zero means no connection at all, one means the two move in lockstep. So the pattern of use tracks distress more than twice as strongly as the clock does.

How many teenagers fall into the problematic group? A WHO and HBSC study of almost 280,000 schoolchildren aged 11, 13 and 15 across 44 countries and regions in Europe, Central Asia and Canada found that in 2022, 11% showed signs of problematic social media use. Four years earlier it was 7%. Girls 13%, boys 9%.

Eleven percent is a lot, and it is rising. But it is eleven, not one hundred. An age ban hits everyone: the 11% who are genuinely struggling and the other 89% for whom social media is how you stay part of your class. And it regulates who and how much, not how, which means it misses the variable most strongly tied to harm.

One caveat matters here. The meta-analysis authors stress this is association, not proven cause. It may equally be that anxious or depressed teenagers retreat into their phones rather than the other way round. Telling those two explanations apart needs long follow-ups of the same people, and there are still few of those. If you want the physiology of why a feed is so hard to put down, we covered that in a separate piece on dopamine and infinite scroll.

And yet country after country is repeating the Australian move.

Who has banned social media for children: the 2026 map

By autumn 2026 age limits on social media are in force or in preparation on at least four continents. Country by country:

• Denmark - under 15. Political agreement signed on 7 November 2025. Parents will be able to allow access from 13, and age will be checked through the national digital ID (an electronic identity almost every Dane over 13 already has). Around 24.8 million dollars was allocated to child online safety.
• Greece - under 15, planned start on 1 January 2027.
• United Kingdom - restrictions under consultation with spring 2027 in view. The consultation covers not only age but overnight curfews, screen time limits and algorithmic restrictions.
• Australia - under 16, in force since 10 December 2025. The regulator has opened investigations into five platforms.
• Malaysia - under 16. The Online Safety Act 2025 took effect on 1 January 2026 and active enforcement began on 1 June. The rules apply to platforms with at least 8 million users in the country, age is verified through the MyKad national ID, a passport or a digital ID, and fines reach 10 million ringgit, roughly 2.5 million dollars.
• Indonesia - under 16. Ministerial Regulation No. 9 of 2026 has applied since 28 March 2026.
• Kazakhstan - under 16. A bill is before parliament with public backing from the president; age verification remains the sticking point.
• Russia - proposals to bar children under 14 from registering on their own, with identity checked through a state service. There is no single approved model in the public domain yet, only a set of initiatives.

If a ban does not solve the problem, what does? The answer is unexpectedly concrete, and it lives in the kitchen and the bedroom rather than in parliament.

What actually works for parents

Two things track a teenager's screen time more than anything else: where the phone sleeps, and whether a parent knows what their child is watching.

That is not an opinion but an observation. In 2026 the journal Acta Paediatrica published work using the large American ABCD study: 7,947 teenagers, mean age 12.9, followed for a year. Researchers checked which parenting habits predicted a rise or fall in screen time. Some of the data came not from self-reports but straight from an app measuring phone time.

Linked to more screen time:

• Screens in the bedroom. The largest link in the whole study (0.18 on the app-measured data), plus more signs of problematic phone use.
• Screens at the family table. More total screen time and more problematic phone use.
• Screens as reward or punishment. "Behave and you get your phone" raises both time and signs of compulsive use. That makes sense: it turns the screen into the most valuable prize in the house.
• The parent's own phone. The more the adult scrolls, the more the teenager does. The link is weak, but it is there, and it runs both ways.

Linked to less screen time:

• Parental monitoring - the adult knowing which apps their child is in and what happens there. Minus 0.08 on the measured data.
• Clear limits - an agreement about time, rather than a phone confiscated on impulse. Minus 0.07.

These numbers are small, and the authors say so: the effects are weak, some data is self-reported, the follow-up ran only a year and overlapped with the pandemic, and the links could run in either direction. But the direction is consistent across all six habits, and it matches what Australia and the English schools showed: what works is not cutting off access but what happens around the screen.

In practice that turns into a short list.

1. The charger lives outside the bedroom. The single habit with the largest effect in the data. It only works if the rule covers adults too.
2. Dinner without phones. Twenty minutes a day, with no exceptions.
3. The phone is not a prize. Do not confiscate it over grades or hand it back for chores, or it becomes currency.
4. Ask what they are watching, without an interrogation. Monitoring in this research means awareness, not surveillance. A child who is scared to tell you simply moves to an app you have never heard of.
5. Watch the pattern, not the clock. Three hours messaging friends and three hours comparing your body to strangers are not the same three hours. The warning sign is not duration but the inability to stop, secrecy, and a mood that collapses without access.

That last point deserves unpacking. Appearance pressure drives problematic use in teenagers harder than almost anything else, and how that works is covered in our piece on looksmaxxing and the price of a perfect face. The "one more try and I will get lucky" mechanic that powers both feeds and games is unpacked in our article on loot boxes and gambling design.

What happens next

The wave of bans will not stop, but its design will probably change.

Denmark is finishing its age verification app, Greece is preparing a January 2027 start, and the UK is looking at spring 2027 while debating curfews and algorithm settings alongside age. The Australian regulator is investigating five major platforms, and the platforms have their own argument: Meta publicly proposes a single age signal delivered once at the operating system or app store level rather than repeated inside every app.

One honest caveat. The Australian ban was three months old when it was measured and less than a year old by autumn 2026. That is not long. In three years the picture may look different: platforms may learn to check age properly, and not having an account before 16 may harden into a generational habit. Or teenagers may simply finish their move into apps the law never reached, and nobody will notice.

For now the data says one thing: disconnecting an account is easy, changing behaviour is not.

What to do tonight

One thing, two minutes: agree that every phone in the house, yours included, sleeps in the kitchen rather than a bedroom. Move the charger there and say so over dinner.

Australia spent a year, 4.7 million disconnected accounts and the threat of 33 million dollar fines per platform to move the number by four points. A kitchen power socket costs less.

Sources

• The Conversation: analysis of the eSafety report on the ban's first three months (2026)
• eSafety Commissioner, Australia: social media age restrictions
• Meta: compliance with Australia's social media ban (2026)
• SMART Schools, The Lancet Regional Health - Europe (2025)
• European Psychiatry: meta-analysis of 59 studies and 547,697 participants (2026)
• WHO Europe and HBSC: teens, screens and mental health (2024)
• Acta Paediatrica: media parenting practices and adolescent screen use, ABCD data (2026)
• Tech Policy Press: early lessons from Australia's ban for the rest of the world (2026)
• JURIST: Indonesia's Regulation No. 9 of 2026
• Malaysia's Online Safety Act 2025 and enforcement from 1 June 2026
• Biometric Update: Denmark's age checks for under-15s
• TechXplore: Qustodio data on app use by Australian children (August 2026)