Alessandro 'kLeZ' Accardo personal website

The personal website of an Italian developer. /nick kLeZ, since forever.

Social media considered harmful. Maybe... (3/5): How to survive

Posted on • by

In the first two instalments I explained that they knew, and that those who tried to stop them didn’t get very far: the odd jury, yes, laws hardly ever. Meanwhile, as the courts take their course and parliaments debate, social media stay exactly what they are, and they sit in the pockets of parents and children, tonight just like last night.

This is the most practical instalment of the series. I gathered what paediatricians, psychologists and health authorities recommend, and for each piece of advice I went looking for evidence that it actually works, because there’s plenty of common sense going around and a lot less data. Some things work, if only a little, and all the work is left to the people who use them and to families, which is, in my view, the most serious part of the whole business.

§At what age

The question every parent asks sooner or later gets different answers depending on who you ask, and put side by side they already tell you quite a lot:

Who says it Smartphone Social media Other
Italian Society of Paediatrics, 2025-2026 not before 13 as late as possible, even after the legal age no screens under 2; less than an hour a day between 2 and 5; less than two after 5
French “Enfants et écrans” commission, 2024 no phone before 11; between 11 and 13 only without internet not before 15, and only “ethical” social media no screens under 3
Proposed European Kids Act, 2026 - none under 13; between 13 and 15 a “mini account” with one hour a day; from 15 a full account still to be approved
Australian law, 2025 - 16 badly enforced, see the previous instalment
United Kingdom, ban announced in 2026 - 16 in force from spring 2027; Ofcom already warns it will be circumvented
Italian law - under 14 parental consent is required the bill on age 15 is stuck in the Senate
The platforms themselves - 13 age chosen to comply with an American law on children’s data

The last row is the one that actually decides things, and it has a rather unedifying history. That 13 comes from a 1998 American law that bans collecting personal data from children under 13 without parental consent, and the platforms wrote it into their terms of service so they wouldn’t have to think about the problem. It’s a threshold born for targeted advertising, and for years Instagram didn’t even bother asking new users for their date of birth: it only started at the end of 2019.

Italian law too, if you look closely, starts from data: under 14 a young person cannot consent on their own to the processing of their data for an online service, and parental consent is required, assuming the platform can be bothered to check it.

I, as you’ll have gathered, side with the strictest row, and I still think we should do everything possible to keep kids away from these platforms for as long as possible. But the data from the last few weeks says that bans, the way we’re writing them, are excellent statements of intent and work poorly in practice. The most instructive case is the United Kingdom. Since July 2025 the Online Safety Act has required platforms to protect minors from harmful content and to verify their age, and on 22 October Ofcom, the regulator in charge of enforcing it, published its first annual report: in the following nine months the share of young people coming across harmful content did not change significantly, and most British teenagers say they saw at least one such item in the space of four weeks. The same report shows where the checks worked and where they didn’t: on porn sites young people’s access dropped, to the point that half of those who visited one only went through sites with age checks, while the pornography that arrives through social media, chats and games stayed exactly where it was, and it mostly comes from personalised feeds. Checks work where the harmful content is the product, and fail where it’s the algorithm delivering it. How exactly is explained by research commissioned by the UK government and published in July: 39% of 11- to 17-year-olds say they successfully got round at least one age check, almost always by lying about their date of birth, and one in ten did it by changing their apparent location, usually with a VPN, which more than one in five teenagers used in the last three months. When you ban something nationally, smuggling springs up, and in the digital world smuggling costs as much as installing an app, often free or a few euros a month at most. The figure that says more than the others, though, is that 43% of teenagers using a VPN say a parent pays for it. So the ban runs up against technology, which makes it easy to get round, and against a culture that no law can change, especially within families. And meanwhile London is preparing to ban social media for under-16s from spring 2027, while Ofcom already warns that kids will find a way round that ban too.

As long as things stay this way, the decision rests on families’ shoulders, and on this the Italian paediatricians have been as clear as anyone can be: first smartphone at 13, and social media as late as possible.

§At home

In November 2025 the Italian Society of Paediatrics presented its new recommendations to the Senate, the result of a review of 78 studies selected from more than 6,800, and summed them up in seven words that all start with P: postpone, protect, preserve adult presence, put clear rules in place, participate in digital education, prevent risks and addictions, promote real experiences. It sounds like a slogan, and partly it is, but a slogan a tired parent remembers at nine in the evening is worth more than a report nobody will ever open. In the press release there’s also a sentence I really like, and which in my view holds the whole argument together better than the seven Ps:

Every year gained without digital is an investment in mental, emotional and relational health.

Behind that sentence there are numbers that justify it, and they’re striking especially when it comes to the youngest: under two, an extra half hour of screen a day can double the risk of language delay, and between 3 and 5 every extra hour of screen takes away about a quarter of an hour of sleep. For teenagers, more than two hours a day in front of a screen is associated with a 67% higher risk of being overweight. Elena Bozzola, who coordinated the work, lined up what paediatricians see in teenagers, namely anxiety, isolation, social media addiction and loss of self-esteem, and added, rightly, that there’s no need to demonise technology.

On this point the American Psychological Association goes even further: social media, it writes, are in themselves neither beneficial nor harmful, and they can give a young person support, companionship and intimacy. That’s why it recommends teaching kids how to use them before handing over the phone, the way you teach a child to ride a bike before letting them out on the road, and adapting features and permissions to age, because an app designed for an adult isn’t made for a twelve-year-old. It’s a less severe position than mine, and yet the two sit together perfectly well, because the problem is how they’re built, and as long as they stay built this way it’s best to get there late, and get there prepared.

As for house rules, the documents I read almost all say the same thing, and they say it with a disarming simplicity. The Italian paediatricians sum them up in a few lines: no phones or tablets in the bedroom, none at the table, none before bed. And among the things that genuinely protect kids they list parental presence, conversation and rules agreed together, as a family, rather than handed down from on high. Then there’s the part no parent wants to hear, which the Italian Society of Paediatrics writes without beating about the bush: the adult remains the role model, and it’s up to them to set the example by using digital in moderation. The society’s president, Rino Agostiniani, put it even more plainly when presenting the recommendations: adults’ presence and example are the first form of digital prevention. And he’s right, because if the first one scrolling the feed at the table is the parent, the no-phones-at-the-table rule lasts exactly one dinner.

I have a daughter. And to my daughter, like many around me (sadly), at first I gave screens. A cartoon while she ate to keep her “calmer”, maybe an hour or so of telly in the afternoon. It doesn’t sound like much, and many of my generation go much further, “but I’m moderate”, I told myself, “I’m a computer scientist, I know the risks of digital and audiovisual entertainment”. But I’m also a careful observer, and I could see the effect on the little one was negative, I could see that the few times there was no way to get a screen she was, yes, harder work for us, but she was more “alive”. I didn’t know how to explain it, it was just a feeling. Then I decided to run an experiment (I like the scientific method), we talked it over and decided to completely eliminate screens. All of them, none excluded, no exceptions. The improvement was clear-cut, her behaviour had improved and so had the way she spoke and expressed herself.
At last I understood, because I’d seen it with my own eyes and experienced it first-hand, the role entertainment content plays on children. We went from a suspected ADHD to a calm, sociable little girl who pays attention to what goes on around her and to what she does at nursery. Now that we’re at primary school I see the difficulties kids face and I can easily weigh them against what I saw in that first year of nursery, before deciding on the total ban, a wise decision with hindsight.
We’ve gone back to having screens on now and then, but in moderation, and I’ve realised that our rules are in fact the same as the SIP’s, even though I’d never read those reports (I’m proud of this, and glad to have caught in time a situation that at primary school would have bordered on the catastrophic).

On two fronts, in Italy, something is moving even without relying on families’ goodwill, though families should still do their bit one way or another. After all, it’s first and foremost in their interest (I speak from direct experience).

The first is school. Since the 2024/25 school year smartphones have been banned in primary and lower secondary schools, and since 2025/26 in upper secondary schools too, for the whole duration of lessons and even for teaching purposes: you can bring the phone, but it stays in your bag. The exceptions are few and sensible, namely students with disabilities who use it for their individual education plan, and IT and telecommunications courses.

The second is the SIM. Since November 2023, under an AGCOM resolution, all operators must offer parental controls for free, and on SIMs registered to a minor they must already be active: only the adult account holder or a parent can switch them off. Some operators weren’t activating them and the Authority had to call them to order, so, if you have a child with a SIM in their name, it’s worth checking in person that the filter is really there.

§Adults fall for it too

Studies and laws focus on minors, because that’s where the harm is scariest, but anyone who thinks this is a kids’ thing, and that adults are safe from it, would do well to read up on one of the most solid studies around.

Between 2004 and 2006 Facebook reached American universities one campus at a time, and three economists had the idea of using that staggered timing as a natural experiment, comparing students at universities where Facebook had already arrived with those where it hadn’t yet. The result, in the figures reported by the Surgeon General, is that after Facebook arrived depression rose by 9% and anxiety by 12% relative to baseline, and the most likely mechanism, according to the authors, is unfavourable comparison with others. They were university students, that is adults, and the Facebook of the time was far simpler than what we carry in our pockets today.

Fortunately there are also studies looking in the opposite direction, measuring what happens when you take it away.

In 2018 a team at the University of Pennsylvania split 143 students into two groups: one carried on as usual, the other for three weeks used Facebook, Instagram and Snapchat for at most ten minutes a day each, with usage checked via the phone’s battery screens and so without trusting what participants reported. In the second group loneliness and depression fell significantly.

In 2020 four economists, from Stanford and New York University, published an experiment run in the four weeks before the 2018 American midterm elections. They recruited almost 2,700 Facebook users and, among those willing to deactivate it for less than 102 dollars, paid a randomly chosen portion to actually switch it off. Those who did found themselves with roughly an extra hour a day and spent it offline, partly in front of the television and partly with friends and family. Happiness and life satisfaction improved a bit, anxiety fell a bit, and so did political polarisation, at the cost of knowing slightly less about the news. The finding I find most interesting, though, comes once the experiment was over: those who had stopped kept using it less, and the value they placed on Facebook had dropped, as happens with certain habits that, once broken, you find you miss less than you thought.

These are small effects, and none of these studies promises a new life, rainbows and pink kangaroos. They say that using them less makes you feel a bit better, and that quitting for a month is less hard than it seems, which, for a habit that’s been tailored to us with such care, is no small thing.

I left every social network before the pandemic, and facebook in particular since at least 2016 (maybe ‘17, I don’t remember my last post). I don’t miss it at all, and in fact I have much more time for myself, my family and my passions (writing among them). An important point is precisely the last one, writing. Since I stopped using Facebook I’ve taken up this positive hobby again, and if you’re here reading something right now it’s precisely because of that decision, made partly at random and partly on the gut feeling of someone who realised something wasn’t right. The last one to go, Twitter, I only kept to grow the blog’s readership, so all things considered I made positive use of it.

§What works, all told

From the studies I found, a few things emerge that you can do straight away, even tonight, almost all with evidence behind them.

On notifications, the idea that springs to everyone’s mind, switching them off, isn’t the best one. In 2019 a team of researchers from Duke University and the University of British Columbia split 237 people into three groups: normal notifications, notifications batched and delivered three times a day, no notifications. The three-deliveries group felt more attentive, more productive, in a better mood and less stressed, while the no-notifications group reported more anxiety and more fear of missing out. Apparently, switching everything off makes you want to check constantly, and batching works better: iPhone even has a dedicated feature for it, the scheduled summary.

Then there’s black and white, which looks like a cheap trick and actually does something. A study published in 2023 had 84 people use their phone normally for one week and in greyscale the following week: screen time dropped by twenty minutes a day, and people felt more in control of their use and less stressed. The number of times they unlocked the phone, though, stayed the same, which means you keep checking it as before but linger less. The reason is intuitive: a colourful app is designed to catch the eye, and without colour it catches it a little less.

Who you follow matters too. The American Psychological Association recommends that teenagers limit content that pushes them to compare themselves with others, especially content about physical appearance, and the advice applies just as it is to adults. It fits with the university study, where the harm came precisely through comparison, and unfollowing people who make you feel inadequate is mental hygiene, and you don’t owe anyone an explanation.

The phone-out-of-the-bedroom rule, recommended by the Italian paediatricians, the Surgeon General and the American psychologists, applies to adults too. If the phone stays on the bedside table because it’s your alarm, a real alarm clock costs ten euros and at two in the morning it doesn’t suggest anything to watch.
The healthy alternative is to use the phone’s “Do Not Disturb” (DND) feature, which works well for suppressing notifications and unexpected calls but lets through alarms, close family numbers (favourites on Android) and anyone who calls three times in a row within 15 minutes, which may signal an emergency.

Then there’s a tool everyone already has in their pocket. Both iPhone and Android have a panel showing how much time you spend on each app and letting you set a daily limit (on iPhone it’s called Screen Time, on Android Digital Wellbeing). I didn’t find solid studies on the effect of these limits on their own, so I wouldn’t treat them as a cure, but the weekly number, looked in the face for the first time, makes quite an impression, and it’s a good starting point.

§The settings that already exist

The platforms already have, in their menus, some features that reduce the harm. Some they added because a law forced them to, others under pressure from parents, newspapers and courts, and generally they don’t do much to make them easy to find, because they reduce the time spent on there, which is the metric they make money from and therefore try to maximise.

The most useful one, in Europe, is the non-personalised feed. Since August 2023 the Digital Services Act has required very large platforms to offer, for each recommender system, at least one option not based on profiling. Facebook added a tab with only the posts of people you follow, in chronological order and without suggestions; Instagram lets you see a feed with only the accounts you follow; TikTok lets Europeans switch off the personalised feed, and in its place shows the most popular videos in the user’s area and worldwide. Names and menu paths change often, but in the feed settings the option to look for is usually called “non-personalised”, “following” or “chronological”.

That it helps is shown by a very large experiment, run on Facebook and Instagram during the 2020 American presidential election and published in Science in 2023: those moved to the chronological feed for three months spent much less time on the platforms and did less on them. Political opinions, which were the real question of the study, didn’t change measurably, but time did, and time is exactly what we care about here. The tailor-made feed is where the little machine from the first instalment spins fastest, and in my view turning it off is the most effective thing you can do short of deleting your account.

Then there are teen accounts. In September 2024 Meta launched teen accounts on Instagram: private profile by default, messages only from people the teenager already follows, sensitive content filtered, a reminder after sixty minutes and a night mode, from 10 pm to 7 am, with notifications muted and auto-replies to messages. Under 16, loosening the restrictions requires a parent’s permission, and on paper the list is good, reassuring even.

A year later that paper was put to the test by Arturo Béjar, who at Meta worked precisely on safety: he tested 47 of the 53 tools the company lists on its website, using test accounts registered to fake teenagers, and a group of researchers from New York University and Northeastern then repeated the tests. According to their report, two tools out of three don’t work or no longer exist. The test accounts still reached content on self-harm and eating disorders, search suggested alternative spellings for blocked words, and the platform recommended following unknown adults. Meta replied that the report misrepresents its commitment to giving parents tools and protecting teenagers (“misrepresents our efforts to empower parents and protect teens”), and that those using these accounts see less sensitive content, receive less unwanted contact and spend less time on Instagram at night.

The report comes from people calling for more regulation, but the tests are described and anyone can repeat them. The practical lesson, for anyone with a child on Instagram, is simple and a bit bitter: the “teen account” label isn’t a guarantee, and it’s no substitute for a parent keeping an eye.

In the 17-billion settlement I talked about in the previous instalment, Meta commits to giving American minors a two-hour daily limit, a night-time block, hidden likes and a non-personalised feed. That means these features can be built, and partly already exist: what was missing was the will to switch them on for everyone. If they exist for an American sixteen-year-old they can exist for anyone, adults included, as the default setting. The European guidelines on the protection of minors, published by the Commission in July 2025, already ask for something similar: autoplay, push notifications, read receipts and streaks of consecutive days not to be broken switched off by default, and an algorithm that gives more weight to what the user says they want than to what they do with their thumb. For now they’re guidance, which the Commission uses as a yardstick when checking compliance with the regulation, and it would be the Kids Act that turns them into written obligations.

§When you need help

In the first instalment I quoted the figure from the Istituto superiore di sanità: in Italy 13.5% of teenagers, that is more than one in eight and almost three in a class of twenty, use social media problematically, with the typical signs of addiction, against a European average of 11%. The same study, at European level, draws a clear line between those who use social media a lot and those who use them problematically, and the distinction matters, because teenagers who use them heavily but without signs of addiction say they feel more supported by their peers. More than the hours, in short, what matters is what happens when you try to stop. If you recognise yourself in those signs, or recognise them in your child (not being able to stop when you’d like to, feeling bad when you can’t use them, the rest of life taking a back seat), know that it isn’t a matter of willpower, and that there’s nothing to be ashamed of in talking to someone about it.

Since 24 June this year the Istituto superiore di sanità has had a national freephone number for all addictions, free and anonymous, Monday to Friday: 800 940 789. The menu asks for the area, and among the areas are new technologies. The first step can also be the paediatrician or the GP, and for paediatricians the Institute has prepared a dedicated guide on behavioural addictions in young people.

§All of this falls on the users, and that’s the problem

Rereading the advice in this instalment, namely putting off the first phone, batching notifications, switching off the personalised feed and the colours, choosing carefully who to follow, checking that the filters are really there, you notice that each one puts the work on the shoulders of whoever uses social media, or whoever in the family bears responsibility for them, and relies on their ability to resist an app that thousands of engineers, paid precisely for this, designed to beat exactly that ability.

The US Surgeon General wrote it too, in the same document I took the numbers for the first instalment from:

The onus of mitigating the potential harms of social media should not be placed solely on the shoulders of parents and caregivers.

And the European Commission, in its preliminary findings on Instagram and Facebook, said the same thing from another angle: parental controls aren’t enough precisely because they offload the work onto parents.

So yes, the things I’ve written should be done, because they help, but knowing that you’re playing defence against an opponent that earns from every minute you give it and has more money and more engineers than any family.

And it’s worth knowing why that work falls to families, too. Institutions have done something, and in these instalments I’ve given them credit for it: Australia had the courage to ban, Italy took phones out of the classroom, the European Commission put it in black and white that parental controls aren’t enough. But they stood by and watched for too long, and when they moved they did so late and with a light touch. Facebook is more than twenty years old, and Europe has only required a feed not based on profiling since 2023. In the United States the law that shields the platforms has been there since 1996, and the one on children’s online safety, after two years of back and forth between Senate and House, still isn’t law. In Europe the independent audits skipped precisely the articles on harms, and the rules on design for minors are still guidelines, awaiting a Kids Act that’s yet to be debated. These are lukewarm solutions, arrived late, that skirt around the business model without touching it, and whoever chooses solutions like these leaves the bill to families, one by one, at the dinner table in the evening.

The real game is won by changing the pitch, that is the way a social network is built and the way it makes money.

In the next instalment I’ll try to do exactly that: describe what a social network that doesn’t need to keep you there would look like, from the user’s point of view.

§Sources