When discussing Internet regulation, one often hears the trope that the Internet must be made "as safe as the offline world".

Indeed, the outgoing Information Commissioner, Elizabeth Denham, gave a speech to the Oxford Internet Institute last week, in which she said:

The internet was not designed for children, but we know the benefits of children going online. We have protections and rules for kids in the offline world – but they haven’t been translated to the online world.

This is part of the broader "Wild West Web" fallacy — that the Internet is unregulated, at odds with the "real" world.

So just how safe are children in the "offline world"?

(As an aside, the Information Commissioner's role relates to the enforcement of the UK's data protection, privacy, and information law frameworks. It's not a general "online safety" or "child protection" role.)

A trip to the playground

The ICO asserts that its "Age Appropriate Design Code" is:

not restricted to services specifically directed at children.

In other words, the ICO considers that its code of practice should apply anywhere a child is likely to go online, whether the service is aimed at children or not.

Internet-accessed services, in its view, should be made "safe for children" (whatever that might mean — see below) by default.

So let's take a quick offline trip to the playground, from the perspective of an unsupervised child running out of a nearby home, off to play on the toys on a lovely summer's morning.

The playground: a place directed at children

I've picked the example of a playground for a reason: because it's a place designed specifically for children. In other words, a place where — presumably — there should be the highest possible standards for child protection.

If the offline world does not have particularly robust protections for a child going to play at a place specifically directed at them, I'm sceptical that there's any truth to the proposition that the offline world as a whole is safe for children by default.

(I took most of these photos in mid-2019, and nothing has changed since then in terms of the issues I discuss.)

Supervised or unsupervised?

I'm using the example of an unsupervised child, gleefully rushing out to play in their nearby playground on a lovely summer's day, without a care in the world.

If the child was supervised by their responsible adult, the risk profile likely looks very different. But all the calls for technical measures online that I have seen are predicated on protecting children who are using the Internet unsupervised and unaccompanied, and so, for the sake of equivalence, so is the fictional child here.

If the offline world is only safe by default for supervised and accompanied children, to argue that the online world must be safe for unsupervised and unaccompanied children is to adopt a completely different standard, eliminating fair comparison.

Different child, different risks

Children are not homogenous. Clearly, "childhood" covers a significant range of ages, and, most likely, both physical and mental abilities, including coping mechanisms and sensitivities.

Some of the risks below are more likely to result in harm to younger children than to older children, but the key thing is that none of them are protected beyond their own education and training.

So here we go: an unsupervised child heads off to the playground

The child dashes out of their bedroom, and falls down the stairs

The child jumps out of bed, sees the sun shining brightly through their window, and they decide it's the perfect morning to go and play outside.

They rush out of the bedroom, and head straight for the stairs.

Do dwellings have stair gates fitted by default? I've never seen one. Teaching children to safely ascend and descend the stairs, accompanying them as they are learning, and supervising them when they initially assert their independence, is part of the role of their responsible adult.

They've hardly taken a step from their bedroom door, and they're already faced with the first reality check that the offline world is not safe for children by default.

Of course, their responsible adult might have had the foresight to install a stair gate, to prevent unsupervised, unaccompanied use. One would hope so, if the child was at risk of falling down the stairs. But the UK's online harms proposals, and the ICO's age appropriate design code, focus only on the role of the service provider, not of the responsible adult. And, in any case, stair gates cost money, and need to be fitted correctly.

They open the front door easily...

Unless you put the key in the lock and lock it, anyone can open the door from the inside just by pulling down the handle. There's no chain.

Convenient in an emergency, and equally convenient for a child making a bid for freedom.

So, just a quick, easy manoevure, and out they go.

Assuming, that is, they don't get their fingers trapped between door and jamb as it shuts, since there's no soft closer fitted, nor a hinge guard.

... and run into the road

Road owners are not required to ensure that children cannot step onto a road and, in practice, the vast majority of roads are but a single step away from the pavement. Here, a child could be out the front door and on a road in front of a vehicle in seconds.

Nor, aside from a slight kerb, is there anything physical to stop a car from riding up from the road onto the pavement. So even if the child walked only on the pavement, the only thing stopping them from being hit is a driver's own self-control.

Even at pedestrian crossings — places intended for people to cross roads, and where people will conceivably be waiting for traffic to step — there is no requirement of a barrier or other restriction. Indeed, there's nothing physical to stop a driver from deciding not to stop, or just failing to notice that they need to stop.

A "safety by design" principle would demand physically-protected crossings, which could be raised by drivers on demand, provided that no pedestrians needed to cross. In other words, a reversal of reality.

But there's no pedestrian crossing here anyway, so it's moot — a child would need to cross the road without a dedicated crossing point, even though it's a common route to nearby primary and secondary schools, as well as the playground.

They have to walk on a road to get to the playground

It's not just walking across a road. There aren't even safe walking routes alongside all roads.

To get to the playground, a child here would need to walk on the road, which turns — at some unmarked point — into a path. A path which is also a cycle path.

Or, worse, coming the other way, the path turns, at some unmarked point, into a road.

The narrow path to the bridge is lined with stinging nettles

So if a cyclist is riding towards them and shows no signs of stopping, the child has a chioce: get hit, or jump into the nettles.

(As an aside, the council has thoughtfully provided numerous cyclepaths in the area. Separating cyclists from vehicles and from pedestrians is safest for everyone. Such a shame, then, that the cyclepaths I would otherwise use tend to be littered with parked cars.)

The bridge over the deep water has railings. The river bank does not

The railings would make it pretty tricky for a child to fall into the water from the bridge.

But the railings are only on the bridge. The bank of the river is left entirely unprotected, and within dashing distance of the playground.

There's even a convenient worn-away path in the bank.

A playground, designed specifically for children. But only under adult supervision

If they've made it as far as the playground, they are greeted with a sign.

A sign which says:

Children should be supervised by an adult.

An area designed for children to play is not safe for children to use on their own. They should be supervised by their responsible adult, who can spot risks, and ensure that they have a safe (and hopefully fun) visit.

If a playground requires adult supervision to be safe, do we really expect the web to be any different?

And note that it is just a sign. It does not stop unsupervised use. It's just a sign on a gate.

And for whom is the sign written? The absent adult? Presumably not for a child, given the complicated grammatical construction of "should be supervised by", rather than something more likely to be understood by a child (such as "wait outside until your adult can watch you play safely").

But why is the sign even necessary? Surely a child's playground is very safe.

Probably because of the numerous heights from which a child could fall or the four swings they could run in front of.

And the protection measures that are there? Some wood chips, and some rubber matting. These are measures designed to protect children from the hazards presented by the equipment itself — from falling, or landing badly — not protecting a child from the actions of someone else in the playground, such as bullying or abuse or assault.

(As an aside, when I was last there with some friends and their children, I spent time picking numerous shards of broken glass out of the wood chips, where someone had smashed a couple of glass bottles against the metal equipment. Is that an environment safe for children?)

And if the playground is aimed at children with a minimum age of 8, why are some of the swings — those with the red superstructure in the photo — seemingly aimed at much younger children?

Note that the sign says that the playground is intended for children between 8 and 13. And note also the absence of any form of age verification: there's nothing to stop younger, or older, children from playing there.

Yes, there is a fence around it. But, helpfully, there are multiple gates, which are easy to open, and which are helpfully painted in a different colour. One of those gates opens within a few metres of a well-used cycle path, and other is but a short dash away from the unprotected bank of a river. And the cycle path.

Anyone could be there

If you're an adult, you don't need to verify that you are there with a child. Any adult can enter the playground. Any adult could talk to a child, or offer sweets.

Or loiter nearby, waiting for the right time. Perhaps sitting on one of the conveniently-placed benches, either inside the playground, or else right by the fence.

And the fence is low enough that someone could easily reach over and grab a child, before heading to the car parking spaces around less than 100m away.

There's no constantly-monitored CCTV, operated and staffed by the playground provider, observing every child's actions.

Additional safeguards during coronavirus

The playground was — like all playgrounds — closed in April 2020, during the coronavirus pandemic.

The effective technical prevention measure, to inhibit potentially lethal usage? About 2 metres of red and white tape, and a small orange notice.

Since Denham talked about "rules" to protect children, presumably this sufficed, since the prohibition on use was enshrined in statute, a Parliament-made rule.

Protecting children using Internet-accessed services

This blogpost is addressed solely at the fallacy that the offline world is safe for children by default, and that the online world should replicate the offline world.

Of course we should look at how to protect vulnerable members of society, whether offline or online. The question is how we do so in a proportionate way, against reasonably foreseeable risks.

If anything, perhaps this post shows that we need to make the offline world safer for children, and indeed other vulnerable inhabitants too. Even offline places dedicated to children are not free of potentially significant risks of harm.

This post also — intentionally — does not get into the numerous resources available to responsible adults to identify risks against which they wish to protect their children, or to help teach their children about how to travel the online world safely.

Nor does it get into technical risk mitigation measures, although — again — there are many. Heck, I use some of the measures which have come in for criticism recently — VPNs, and DNS over https — to maximise the scope of the filtering of Internet connections. More filtering, and more aggressive filtering, not less.

Indeed, I suspect that it is easier to prevent an unsupervised child from travelling to a particular place online than it is offline, if that's the path down which the responsible adult wishes to go.

If a responsible adult would not let their child travel, alone and unsupervised, around the offline world, until they were comfortable that they understood the risks and had the tools and techniques and support network suitable for their needs and capabilities available to look after themselves, perhaps we should take the same approach when it comes to travelling online, rather than asserting that we must make the online world suitable for unsupervised, unaccompanied access for everyone from toddlers upwards.

But perhaps we can proceed with a discourse on appropriate and proportionate online safety measures, without the fallacy that the offline world is safe for children by default and by design.

Further reading about "online harms"

If you've got this far, you might like:

• this post about the extra-territorial over-reach of the current "online harms" proposals
• this one about the increasing state use of Internet access providers as enforcers, through blocking measures