Digital hygiene for the family: screens without bans, without giving up

Seven in the evening. A kid is sitting at a tablet for the second time this afternoon — the first time because it was genuinely useful, dinner needed finishing, the second time because the first shutdown ended in tears and nobody has the energy for round two. A parent sits across the room answering work messages that can't wait. Both have a screen, both have a good reason, and both would describe this moment as an exception.

A month later, the exception is just what the evening looks like. Then comes the moment nearly every parent recognizes: a decision that this has to stop, followed by a set of strict rules that last nine days. On day nine there's an illness, a work crisis, or a long car ride, and the rules dissolve — this time with a side of failure. From there, two familiar paths open up: either an even stricter ban that produces even more pressure, or quiet resignation, where things are simply left as they are.

Screens are a topic where neither bans nor resignation work — and, more importantly, it isn't really a topic about children, it's about the whole household. Bans run into the fact that devices today are the road to school, to friends, to entertainment and to information, and a child who's forbidden them at home will encounter them everywhere else anyway, just without a guide. Resignation runs into the fact that the environment is designed to hold attention as long as possible, and a child has no more defense against it than an adult does. This chapter doesn't prescribe hour counts or specific age cutoffs; those vary between institutions' recommendations, change over time, and can't be applied identically to every child. It describes mechanics that work in any family — building on the chapter The household as an operation, which covers dividing areas of responsibility, and on Parents and time, which covers time and presence.

Between banning and giving up

Why bans fail so reliably isn't a mystery. First, they're unenforceable: a parent can't be everywhere, and the older a child gets, the more space exists where different rules apply — at a friend's house, at school, at an activity, on the bus. Second, a ban creates pressure that looks for a release valve; something unavailable becomes more valuable, and getting it in secret becomes an accomplishment. And third, and most important, a ban teaches nothing. A child who simply has the device withheld learns nothing about managing their own attention — exactly the skill they'll need the moment nobody is standing over them anymore.

Resignation fails from the other end. “There's no point, I can't win this” is an understandable conclusion after a series of failed attempts, but it assumes the only alternative to full control is none at all. It isn't. In between lies a whole range of things that actually can be maintained: where devices get used, when they don't, what happens on them, and whether any of it gets talked about.

It's worth taking one uncomfortable asymmetry seriously. On the other side of the screen isn't just content — it's design whose explicit goal is to hold the user as long as possible: infinite scroll with no natural stopping point, autoplay for the next video, notifications timed to bring someone back, rewards delivered unpredictably, which is exactly the pattern that learning psychology has shown to be the most habit-forming of all. This isn't a conspiracy, it's a business model. But it means that a failure of self-control is not proof of weak willpower — not in a child, and not in an adult. Anyone expecting a child to resist something they themselves can't resist is building on shaky ground. The chapter Digital minimalism covers this mechanism in more detail.

That leads to the first practical shift. Instead of relying on willpower in the moment — your own or a child's — it pays to change the environment instead, because the environment works on its own and doesn't require a fresh argument every time. Most of the rest of this chapter is about exactly that.

Age and development: there's no single number for everyone

The question “how many hours a day is okay” is the most common one and the least useful one. Recommendations from professional bodies vary between countries, change over time, and above all work with an average that doesn't fit any particular child. Two hours spent making a video together and two hours of endless feed scrolling have nothing in common, even though they log identically on a chart.

It's more useful to look at what's actually happening at a given age, and what a screen either supports or displaces during that period. A parent can translate this lens onto their own child, which a number from a table simply can't do.

That progression isn't a manual or a schedule. It's a reminder that the same rule means something completely different at six and at fourteen, and that rules have to age along with the child. The most common mistake isn't being too strict or too lenient — it's holding onto a rule set at age eight when the child is thirteen. At that point it stops protecting and starts becoming a reason to work around it.

Two things are worth tracking more than time, at every age. The first is sleep: late device use pushes back falling asleep, and a sleep-deprived child struggles with focus, mood and learning. This is the one area where, according to repeated findings, the relationship is strongest and where it's worth being consistent. The second is displacement: what specifically isn't happening during that time — movement, sleep, friends, the boredom that gives rise to play. When none of that is being displaced, the question of hours matters a lot less than it sounds like it should.

Rules apply to everyone, adults included

There's one sentence that can tear down even the best-built system: “I need it for work.” Sometimes it's true. But a child doesn't hear the argument, they hear the exception — and an exception that only one side gets turns the rule into a tool of power. From that moment on, screens aren't negotiated as a shared arrangement anymore; they're something the stronger party extracted from the weaker one.

The fix isn't for adults to stop working. It's for the rule to be framed as a household rule, not a rule for kids — and for the exception to be named as an exception. “No phones at the table” applies to everyone. If you genuinely need to handle something on a Sunday evening, say so out loud: “This is a work thing I have to finish, I'm stepping out with it.” A sentence that flags the exception keeps it an exception. Without one, it quietly becomes the new normal — the exact mechanism by which boundaries at work dissolve too, as the chapter Client-facing work describes.

There's a second, less comfortable reason. Children learn to manage attention mainly by imitation, and what they see is stronger than what they hear. A parent sitting at dinner with a phone in hand teaches something different in practice than what they say out loud. That's not a reproach — it's actually a relief, because it means the most effective intervention into a child's digital habits is an intervention into your own. And that one can be done without a single conflict.

The practical minimum for an adult has three points and takes a few minutes to set up. Limit notifications to things that genuinely need attention right now — most don't. Put the phone out of reach during moments meant to belong to the family; face-down on the table is still within reach. And reduce the device's own pull, which the tip Grayscale display: a phone without color is about half as tempting helps with most cheaply. When something stronger is needed, system-level app limits exist, covered in the tip App limits — and they work just as well on an adult as on a child, which is worth saying out loud.

Device-free places and times instead of counting hours

A time limit has three flaws that make it the worst tool available. It requires ongoing measurement and enforcement, which is extra work for the adult. It creates negotiation over every single minute, because a limit is always contestable. And it doesn't distinguish quality: half an hour of creating something together and half an hour of aimless scrolling burn the same quota.

A different principle works considerably better: place and time are more legible to a child than quantity. A rule tied to a situation doesn't need to be measured, doesn't need to be negotiated, and enforces itself. It doesn't take many — two or three that genuinely hold do more than ten that always end up bent.

• Places. Most often the dinner table and the bedroom. The rule that devices don't sleep in the bedroom, in particular, handles the hardest part of the whole topic — evening use and sleep — without a single argument about time. A shared charging spot in the hallway is a boring measure with a disproportionately large effect, and it's good practice to have the adults' phones go there in the evening too.
• Times. Before leaving in the morning, the last hour before bed, meals together, the drive in the car when there's conversation happening. It isn't about the length — it's that these are moments where presence actually means something.
• Situations. When someone's talking with someone. When there's a guest. When you're heading out. These are rules that are easy to explain, because the reason is visible.

What's left is the hardest thing, and the one talked about least: transitions. Most conflicts around screens don't come from using them — they come from ending them, from something being cut off in the middle. A few mechanical things help. Tie the ending to content, not the clock: “after this episode,” “after you finish the level” is achievable; “in five minutes” in the middle of a game isn't. Give advance notice. Have what comes next already lined up, because the transition from something engaging into nothing is the hardest one there is. And don't use screen time as a reward or punishment — that turns it into a currency, and its value only grows every time it's used as payment.

Content matters more than screen time

If a parent were to take away just one thing from this whole chapter, it would probably be this: the question “what are they doing on there” matters more than “how long are they on there.” A device isn't one activity — it's a space where anything from composing music to endless short-video scrolling can happen, and the difference between those matters more than any difference in minutes.

Two axes are enough to bring order to it: whether the child is creating or consuming, and whether they're doing it alone or with someone.

That matrix isn't a judgment of the child, and the bottom-right quadrant isn't a sin — moments of mindless downtime belong in adult life too. It's more of a map of where to ask questions. When it turns out that almost all the time sits in one corner, that's more useful information than any hour count, and more importantly, something can actually be done with it: offer an alternative within the same environment instead of removing the environment altogether.

There's one more layer to content that gets forgotten in debates about time — what a child runs into without looking for it. Setting up kids' modes and filters makes sense for younger children and is an afternoon's work, but no filter is fully reliable, and its usefulness declines with age. The one thing that works consistently is a child knowing that if they run into something unpleasant, they can say so and it won't turn into trouble. A parent who reacts to the first such admission by confiscating the device won't hear about the second one. This is the single most valuable thing to have on this whole topic, and it takes a long time to build.

AI and kids: what to address before they ask

Language models have brought a question into households that nobody has years of experience with yet. Most services have their own age terms, and some require a parent's consent or involvement — so the first step is finding out what a specific service actually permits for children, rather than assuming. Age appropriateness isn't a formality: for younger children it's usually sensible to work with a model together, and for older ones, independently but with clear ground rules.

Three things are worth teaching before the mechanics of how to use it.

The model is confidently wrong. It doesn't say “I don't know” as often as it should, and a made-up fact looks exactly like a correct one. A child won't remember this from being told — they'll remember it from one experience of their own. Let them ask about something they already know well and find the mistake in the answer. That one moment does more than ten warnings, and it's the foundation of a principle that applies to adults too: AI proposes, a person approves.

Private things don't belong in the model. Full name, address, school, photos, friends' names, anything about health. The explanation doesn't have to be complicated: whatever gets typed leaves our device, and we don't know where it ends up. The same rule, incidentally, applies to adults at work — sensitive data belongs only in a paid account with contractually protected data handling, and even there only after real consideration.

Getting it done isn't the same as knowing how to do it. This is usually where the most tension shows up between home and school. A useful distinction a child can understand: the model is good for explaining, for trying things out, for feedback and for ideas — not for handing in the work in the child's place. And above all, the school's rules apply, even when they don't make sense to parents at home. When it's hard for a parent to get their bearings on this, the view from the other side of the desk in the tip AI in the classroom: teaching students with ChatGPT in their pocket is useful — it shows why teachers set certain limits and what they're actually asking of kids.

One more thing remains, the one talked about least and likely to matter most. Models can respond gently, patiently and without judgment, which can be a relief for a child too embarrassed to ask a person. But it isn't a relationship, and it isn't anyone who bears responsibility for the child. It's worth saying out loud — without making a big deal of it — that for things that hurt, or that a child doesn't know who to tell, there's a parent, and if not a parent, some other adult.

A child's digital footprint, and deciding on their behalf

There's a decision a parent makes on a child's behalf without usually thinking of it as a decision at all: what appears about them online before they can say anything about it. Photos from the delivery room, from first grade, from vacations, stories that are adorable at six and mortifying at fourteen. None of it comes from bad intent — it's ordinary joy and the urge to share it.

It helps to hold a few things in mind without turning it into paranoia. Shared content can't be reliably taken back. Privacy settings change over time, and platforms decide them, not you. The combination of a name, a school, an activity and regular times is information nobody would voluntarily post on a fence about a child, even though each individual detail looks harmless on its own. And the child has their own stake in how they're portrayed, and will eventually start asserting it.

The practical guideline is simple and requires no system at all. Ask the child as soon as they're able to answer — and respect a no, even when the reason is hard to follow. Don't share situations where the child is vulnerable, undressed, sick or crying. Think about traceability: who will see this, can it be used to figure out where the child usually is. And ask the same of others — grandparents, other parents, the school. “Please don't post this anywhere” is a legitimate request, and most people accept it without friction when it's said in advance.

The mirror image of this is a child's privacy from parents. Tracking apps, reading messages and checking browsing history are tools that have their place, but they carry a cost — and that cost is trust. A reasonable stance can be summed up as: transparency instead of secret surveillance. When something is being monitored, the child knows about it and knows why. For younger children, oversight is a given; with age, it should shrink and turn into conversation. And if a serious suspicion arises that something dangerous is happening, checking is warranted — but even then it's better to disclose it than to hide it, because discovered secret surveillance usually ends the one thing that protects a child long-term: that they'll come and say what happened, on their own.

What to take away

• Neither bans nor giving up work. A ban is unenforceable, creates pressure, and teaches nothing; resignation wrongly assumes the only alternative to full control is none at all.
• A failure of self-control isn't weak willpower. The environment is designed to hold attention, and even an adult can't resist it. That's why it's worth changing the environment instead of relying on a decision made in the moment.
• There's no single hour number for everyone. More useful than a limit is asking what a screen is displacing at that age — and watching sleep above all. Rules have to age along with the child.
• Rules apply to the whole household, adults included. An exception has to be named as an exception, or it becomes the new rule. The most effective intervention into a child's habits is an intervention into your own.
• Device-free places and times beat time limits — they don't need measuring, don't need negotiating, and enforce themselves. A device-free bedroom solves the hardest part of the topic without an argument.
• “What are they doing on there” matters more than “how long are they on there.” Creating and watching together are a different thing from an endless feed, even though they count toward the same quota.
• With AI, address the service's age terms, confident mistakes, privacy, and the line between help and handed-in work — and respect the school's rules at home. One firsthand experience of a model's mistake beats ten warnings.
• A child's digital footprint starts with the parent. Ask as soon as the child can answer, and ask the same of everyone around you. Choose transparency over secret monitoring — otherwise you lose the one thing that protects long-term.