At Sixteen, the Algorithm Does Not Become Kinder

On 10 December 2025, Australia became the first country to require major social-media platforms to take reasonable steps to prevent children under sixteen from holding accounts.

Facebook, Instagram, Snapchat, TikTok, YouTube and several other services were included. The policy was presented as a new line around childhood: below it, children would be kept outside; above it, they could enter.

On 15 June 2026, the UK government announced that it would introduce its own under-sixteen ban, with the first restrictions expected to take effect in spring 2027.

The appeal is obvious.

Social media can expose children to bullying, grooming, sexual content, self-harm material, misinformation and products designed to hold their attention for as long as possible. Parents are worried because there are real things to worry about.

If a room is dangerous, locking the door feels like protection.

But a locked door and a safer child are not the same thing.

────────

A Banned Account Is Not an Offline Childhood

The videos do not vanish when an account does.

Neither do group chats, gaming platforms, search engines, shared links or disappearing messages. Other children can still show them things at school. Someone else can still hold up a phone.

The internet is not one room that can be closed.

The contradiction is already visible in how children describe their own lives online.

In 2026, Ofcom found that nearly three-quarters of eleven- to seventeen-year-olds had encountered harmful content. Yet in the same research, 81% of children said they felt safe on social-media and messaging services most of the time, while 71% said they felt happy.

Both findings can be true.

That does not make the harmful content less serious. It shows that children are not entering a place they experience only as danger. The same platform can contain risk, friendship, humour, identity and support.

We cannot regulate ourselves back into a childhood that no longer exists. We have to make children safer in the one they actually have.

────────

My Son and I Speak in Videos

Social media is usually discussed as something children consume independently.

My son and I send each other videos all the time, multiple times a day. Some are funny. Some are about football. Some about people we both know. Some begin a conversation that would not otherwise have happened.

A video is not always content.

Sometimes it is a sentence.

It can mean this made me think of you.

It can mean look at this.

It can mean I do not quite know how to say this, but this person has.

That is not a replacement for conversation. It is one of the forms conversation now takes.

That small exchange between us is part of a much larger way of communicating. When Pew asked American teenagers about their experiences in 2025, nearly three-quarters said social media made them feel more connected to their friends. Almost two-thirds used it as somewhere to be creative. Around half found acceptance or support there during difficult times.

Those figures do not cancel out the harm.

They show why removal also has a cost.

Social media has become part of the language through which children maintain relationships. Taking away the account can also take away part of the conversation.

────────

The Children Who Stop Telling Us

Recently, a video involving an underage child went viral among children at my son’s school.

I will not describe the video. The important part is what happened around it.

Children had viewed it at school, but they did not want to approach the school for help. They believed the response would be tighter policing of their phones and social-media use.

So they held back.

The danger was not only what had appeared on their screens. It was what they believed would happen if they told an adult.

Despite it being an uncomfortable conversation, the boys came home immediately after school to share it with me.

They told me because they knew I understood social media. They also knew that my first response would not be to confiscate phones, limit social media usage or close the conversation.

This one incident does not prove that every ban will fail. It reveals a problem that any safety policy has to take seriously.

If children believe that asking for help will lead to more surveillance, less freedom or the loss of their account, they become more likely to censor what adults see.

The danger does not disappear.

The disclosure does.

In 2025, UNICEF warned that bans used on their own could push children towards less regulated spaces, where they would be harder to protect and less likely to seek help when something went wrong.

That is the difference between making harm invisible to adults and making children safer from it.

────────

The Feed I See Is Not the Feed My Son Sees

When my son and I open the same app, we do not enter the same place.

My feed has learned from the things I watch, the work I do and the subjects that hold my attention. His has been shaped by an entirely different set of pauses, searches, shares and conversations.

Social media no longer has one front page.

It builds a different one around each of us.

That means a screen-time report can tell me how long my son spent on an app. It cannot tell me what version of the internet the app gave him while he was there.

Parents need to see more than the number of minutes.

Sometimes that means sitting beside a child and scrolling with them. Not taking the phone to search for evidence, but seeing what arrives before they have searched for anything at all.

A strange recommendation. An advert made to look like advice. A joke whose meaning exists entirely in the comments. A video that has appeared because they paused on something similar the day before.

TikTok describes its own recommendation system as one shaped by what people watch, like, share, search for and create. The system sees behaviour. It cannot always know whether a child paused because they were interested, shocked or confused.

But it learns from the pause anyway.

Attention becomes an instruction.

One upsetting video can therefore become two. A moment of curiosity can begin to look like a preference. The feed keeps answering a question the child may never have intended to ask.

A parent who understands that can help a child interrupt the pattern. We can ask why they think something appeared, show them how to reject a recommendation, look at the source together or explain why watching something is also teaching the platform what to send next.

This kind of involvement does more than reassure parents. After looking at the online lives of nearly 15,000 children across eleven countries, UNICEF and the London School of Economics found that children with a wider range of online experiences tended to develop stronger digital skills. Those with more limited access tended to have weaker ones.

The important difference was not simply how much access children had. It was whether an adult helped them understand what they encountered. Positive parental involvement improved children’s skills and reduced their exposure to risk.

Not every child has a parent who can do that. Some adults do not understand the platforms themselves. Others are stretched, absent or unaware of what their child is seeing.

That does not make a blanket ban sufficient. It makes digital support a responsibility that must also be carried by schools, platforms and the adults responsible for safeguarding children. A child who cannot find guidance at home still needs somewhere to ask for help without expecting punishment.

Safety is not knowing how many minutes a child spent behind a screen.

It is remaining close enough to understand what the screen showed them, and trusted enough that they tell you when it shows them something it should not.

────────

Sixteen Is an Age, Not a Digital Skill

Rejecting a blanket ban does not mean giving a ten-year-old and a fifteen-year-old the same access. Boundaries should change as a child develops the judgement to manage more freedom.

A ban can keep a child away from a particular account for a period of time.

It cannot teach them how to recognise manipulation, protect their privacy, question a viral claim, understand an advert, leave a hostile group chat or notice when an algorithm is narrowing their view of the world.

At sixteen, the legal gate opens.

The teenager on the other side may have an older face and exactly the same lack of experience.

There are two different kinds of protection here.

Protection by exclusion keeps a child away from a risk while somebody controls the entrance.

Protection by competence gives them knowledge, judgement and an adult they trust when the entrance can no longer be controlled.

We need both.

But only one travels with them.

────────

The Ban Starts at the Wrong End

An age ban changes who may enter. It does not necessarily change what they are entering.

At sixteen, the same endless feed is waiting. The same recommendation systems are learning from every pause. The same public defaults and routes for strangers to make contact may still be there.

Age is the cleanest lever for government to pull. It is also the one that asks least of the product.

Children themselves seem to understand the distinction. When the UK government consulted them in 2026, only 19% supported a minimum age of sixteen across every social-media platform.

That did not mean they wanted adults to do nothing. Among young people aged sixteen to twenty-one, 65% wanted at least some features designed to hold younger users’ attention restricted. More than half singled out infinite scrolling.

They were identifying the parts of the system that needed to change.

That is where regulation should be strongest: private accounts by default, no stranger messages, tighter limits on livestreaming, no overnight notifications, clearer control over recommendation systems, fast child-friendly reporting, proper moderation and consequences for companies that allow harmful material to spread.

Parents cannot individually correct products designed by some of the most powerful companies in the world. Responsibility must remain with the platforms too.

But parents also need to remain close enough to the child’s online life that we are still the people they tell.

────────

The Door Opens Anyway

Social media has already become part of how children discover things, maintain friendships and communicate with their families.

There is no going back to a world in which it was never invented.

The choice is not whether children will enter a digital world.

It is whether they will enter it alone.

When that video travelled around my son’s school, safety did not arrive because the technology had been removed. It arrived when they came home and spoke to me.

That conversation was possible because I understood the world being described.

A locked door can delay the encounter.

It cannot prepare a child for it.

At sixteen, the algorithm does not become kinder.

It’s better they meet it gradually, with us beside them, than meet it fully for the first time after years of learning to hide it.