Children & Technology / Policy Analysis
There is a moment that parents across the world will recognise, the moment a child receives their first smartphone and, within weeks, becomes a stranger who lives in a rectangle of light.
What follows is not a phase. For millions of adolescents, it is the beginning of a years-long neurological and psychological reordering, engineered not by accident but by design, calibrated by some of the most sophisticated behavioral scientists in human history working for companies whose business model depends on capturing as many minutes of a child's waking life as possible.
The debate about whether social media should be banned for children under sixteen is frequently framed as a culture war skirmish, pitting anxious parents and conservative moralists against free speech advocates and digital rights campaigners.
This framing is a distraction. The real question is simpler and more brutal: do we have sufficient evidence that these platforms are causing measurable, lasting harm to children's mental health, and if we do, what does our collective failure to act say about whose interests our institutions actually protect?
The evidence, accumulated over a decade of research, internal corporate documents released under litigation, and some of the largest longitudinal studies ever conducted on adolescent wellbeing, now points in one direction with uncomfortable consistency.
Social media, as it is currently designed and deployed, is making children sicker, lonelier, more anxious, and in measurable numbers, more likely to harm themselves.
Rates of depression among teenage girls roughly tripled in the United States between 2010 and 2019, a period that precisely overlaps with the mass adoption of smartphones and social media among adolescents.
The statistical case begins, appropriately, with the data that should have triggered a policy response a decade ago.
Jonathan Haidt's research, synthesised in his 2024 book "The Anxious Generation," documents that rates of depression, anxiety, self-harm, and suicide among adolescents in the United States, United Kingdom, Canada, and Australia began rising in lockstep around 2012, the precise moment when smartphone ownership crossed the threshold into mass adoption among teenagers.
This is not a correlation that applies to adults, who experienced the same technology adoption but did not show the same mental health deterioration at the same magnitude. It is not a correlation that applies uniformly across genders: girls, who use image-based social platforms at higher rates and are more susceptible to the social comparison mechanisms these platforms exploit, show dramatically worse outcomes than boys.
Between 2010 and 2020, the CDC documented a 153 percent increase in suicide attempts among teenage girls in the United States. Emergency department visits for self-harm among girls aged ten to fourteen increased by 188 percent in the same period. These are not statistical noise.
They are a generation's distress signal, and the algorithm made sure no adult in a position of power received the notification.
What makes the harm not just tragic but morally outrageous is that it was known inside the companies building the platforms long before it was known outside them.
In 2021, the Wall Street Journal published what became known as the Facebook Files, a trove of internal Meta research showing that the company had conducted its own studies on Instagram's effect on teenage girls and found that 32 percent of them said that when they felt bad about their bodies, Instagram made them feel worse.
One internal slide read: "We make body image issues worse for one in three teen girls." The company buried this research. It did not change the product in any meaningful way. It continued to run recommendation algorithms that pushed body image content, diet content, and social comparison content to vulnerable adolescent users because that content generated higher engagement, and engagement is the currency the business model demands.
This is not a corporation that accidentally harmed children. It is a corporation that discovered it was harming children, calculated that the harm was commercially acceptable, and continued.
"We make body image issues worse for one in three teen girls." Internal Meta research, buried and not acted upon for years, while the algorithm kept serving diet and comparison content to adolescent users.
Facebook Files, Wall Street Journal, 2021 — Internal Meta presentation slideThe mechanism of harm is not mysterious once you understand how these platforms are engineered. Social media companies employ behavioral psychologists, neuroscientists, and machine learning engineers whose explicit job is to maximise what the industry calls "engagement," which is simply a sanitised word for addiction.
Variable reward schedules, the same psychological mechanism that makes slot machines compulsive, are built into the like button, the notification badge, and the infinite scroll. When a child posts a photo and waits for likes, their brain releases dopamine in a pattern that is neurologically indistinguishable from the reward cycle of gambling.
The unpredictability of the reward, sometimes lots of likes, sometimes very few, is not a design flaw. It is the design. Pew Research data from 2022 found that 35 percent of American teenagers said they used social media "almost constantly," a figure that should be read not as a testimony to enjoyable products but as a clinical description of compulsive behavior that the adolescent brain, which is not fully developed in its impulse control and reward regulation capacities until the mid-twenties, is structurally incapable of resisting on its own.
Of American teenagers in 2022 reported using social media "almost constantly." The adolescent prefrontal cortex, responsible for impulse control, does not fully mature until the mid-twenties. This is not a coincidence, it is a target demographic. Pew Research Center, 2022
The sleep dimension alone ought to end the debate. Adolescents require between eight and ten hours of sleep for normal brain development, immune function, emotional regulation, and academic performance.
The Sleep Foundation's data shows that more than 70 percent of American teenagers are chronically sleep-deprived, and that social media use after 10pm is one of the strongest predictors of insufficient sleep among this age group. Sleep deprivation in adolescence is not a lifestyle inconvenience.
It is a neurological insult that impairs the very cognitive functions, emotional regulation, risk assessment, and stress resilience, that might otherwise allow a teenager to manage the psychological pressures social media itself creates.
The platforms have in effect engineered a feedback loop: they steal the sleep that would otherwise protect children from the anxiety the platforms generate. The product is both the wound and the removal of the bandage.
Those who oppose an age-based ban typically advance three categories of argument, and each deserves the serious engagement it rarely receives. The first is the digital literacy argument: rather than banning, we should educate children to use social media wisely.
This argument sounds reasonable and collapses on contact with neuroscience. You cannot teach a thirteen-year-old to resist a product engineered by teams of PhDs specifically to override the prefrontal cortex's capacity for rational decision-making. We do not respond to tobacco's harms by teaching teenagers to smoke mindfully.
We do not permit casinos to admit children and then provide gambling literacy classes. The analogy to addictive, commercially engineered products is not rhetorical flourish. It is the most accurate available description of what social media companies have built and what they continue to operate without meaningful regulation.
Increase in emergency department visits for self-harm among girls aged 10 to 14 in the United States between 2010 and 2020 — the decade of mass smartphone and social media adoption. CDC Morbidity and Mortality Weekly Report, 2023
The second argument is the free speech and digital rights argument: restricting minors' access to social media infringes their right to expression, communication, and participation in public life.
This argument, frequently advanced by civil liberties organisations with genuine and important track records on free expression, requires a more careful response. Children's rights to expression and participation are real and must be protected.
But the social media platforms used by most teenagers are not public squares. They are privately owned commercial architectures designed to monetise attention through advertising, to harvest behavioral data at scale, and to maximise time on platform.
Restricting a child's access to Instagram is not equivalent to restricting their access to a library, a newspaper, a telephone, or indeed the open internet. It is restricting their exposure to a specific product whose business model has been demonstrated to cause measurable psychological harm. These are not the same thing, and conflating them serves the platforms' interests more than the children's.
The third argument is the enforcement argument: any ban would be ineffective because children will simply circumvent age verification systems using false information, VPNs, or older siblings' accounts.
This is partially true and entirely insufficient as a reason not to act. The imperfect enforceability of a legal age restriction does not constitute an argument against the restriction. Alcohol purchase ages are imperfectly enforced; we do not therefore remove them. Tobacco regulations are circumvented by some teenagers; we do not therefore permit cigarette vending machines in middle schools.
The purpose of an age restriction is not exclusively or even primarily enforcement in every individual case. It is the establishment of a social and legal norm, the shift of liability from the child to the platform, and the creation of a legal architecture that enables regulators to impose meaningful penalties when companies knowingly allow underage access.
Australia's Online Safety Amendment Act of 2024, which bans social media for under-sixteens and places the compliance burden on the platforms rather than on parents or children, is precisely this kind of regulatory architecture, and its early implementation data warrants serious attention from other governments.
"The purpose of an age restriction is not to achieve perfect compliance.
It is to shift the legal burden from the child to the corporation, and to establish that children are not an acceptable target market for psychologically manipulative commercial products."
Synthesis of regulatory theory — Online Safety Act frameworks, UK and Australia, 2024The Australian case is instructive because it reveals what happens when a government decides to take the public health framing seriously and act accordingly.
The 2024 legislation, passed with bipartisan support after years of advocacy from child welfare organisations and a formal parliamentary inquiry, requires social media platforms to take "reasonable steps" to prevent children under sixteen from creating accounts, with fines of up to AUD 49.5 million for systemic failures in compliance.
Crucially, the law explicitly does not require children or parents to provide identification documents; the responsibility is placed on the platforms to develop age assurance technologies. The platforms, predictably, lobbied against the legislation with considerable force, deploying the enforcement and free speech arguments in the Australian parliament with the same precision they deploy them in every jurisdiction where regulation is proposed.
The Australian parliament, to its credit, declined to be persuaded. The result is the most substantive legislative response to the child social media crisis anywhere in the world.
Maximum fine in Australian dollars imposed on social media platforms for systemic failures to prevent under-16 access, under Australia's landmark 2024 Online Safety Amendment Act, the world's most substantive legislative response to child social media harm. Australian Government Federal Register of Legislation
The United Kingdom has moved along a parallel track. The Online Safety Act of 2023 imposes a duty of care on platforms in respect of child users, requiring age-appropriate design, prohibition of addictive design features for children, and restrictions on the recommendation of harmful content to minors.
Ofcom, the UK's communications regulator, has begun issuing enforcement guidance and has signaled its willingness to use the act's maximum fine of ten percent of global annual turnover, a figure that for a company the size of Meta would represent tens of billions of dollars.
The trajectory in both countries reflects a growing recognition that the self-regulatory model the platforms have operated under for two decades has failed comprehensively and that the framing of child protection as a parental responsibility rather than a corporate liability has allowed the companies most responsible for the harm to escape accountability for it.
The mental health costs being absorbed by healthcare systems that are treating the downstream consequences of social media harm are, on any honest accounting, enormous.
A 2022 Lancet Public Health study found a consistent negative association between social media use and mental wellbeing among adolescents, with the strongest effects in girls and in pre-adolescent users who joined platforms before the age of thirteen.
The cost to national health systems of treating the anxiety disorders, eating disorders, depression, and self-harm resulting from or exacerbated by social media exposure has not been comprehensively calculated, but even a partial accounting would dwarf the revenue generated by these platforms in the markets where the harm is concentrated.
We are, as a matter of public policy, allowing private corporations to export their operational costs onto public health budgets while protecting their right to continue the 2practices that create those costs. This is not an argument the platforms' policy teams engage with in their public communications, because there is no good answer to it.
Of American teenagers are chronically sleep-deprived, with social media use after 10pm identified as one of the strongest predictors. The adolescent brain needs 8 to 10 hours of sleep for normal emotional and cognitive development. Sleep Foundation Research
The counterargument that social media provides community and connection for marginalised teenagers, including LGBTQ+ youth in hostile home environments, is the most genuinely difficult part of the policy debate and deserves serious weight rather than dismissal.
For some adolescents, online spaces offer a form of belonging and affirmation that their immediate physical environment denies them. This is real, and any thoughtful policy response must account for it. But several things are also true simultaneously: the benefit of connection does not neutralise the algorithmic architecture designed to maximise engagement rather than wellbeing; platforms can be regulated to protect minors without eliminating all online community spaces; and the specific features that cause the most documented harm, infinite scroll, algorithmic amplification of comparison content, quantified social validation through public like counts, can be restricted or removed without eliminating digital communication for young people.
The choice is not binary between banning everything and permitting everything. It is between continuing to permit privately designed behavioral manipulation systems to operate on children without constraint, or deciding that children's neurological development is a protected interest that supersedes a corporation's quarterly revenue projection.
"We allowed entire car interiors to be redesigned around passenger safety.
We banned lead in paint because children were eating it. We have never before decided that the profit margin of the manufacturer was a sufficient reason to leave a documented hazard in children's hands."
Analytical synthesis — product safety law parallels in child protection policyThe broader historical context matters here. Every generation has faced a new technology that created legitimate anxiety about its impact on children, and most of those anxieties turned out to be overstated.
Television, video games, the internet itself, all were greeted with moral panics that did not fully materialise into the catastrophic outcomes predicted. This history makes reasonable people cautious about reacting to social media with legislation. But the current situation differs from those earlier anxieties in a critical respect: we have data. We have internal corporate research. We have longitudinal studies. We have emergency department admission records.
We have a documented, quantified deterioration in adolescent mental health that began at a specific moment, corresponds to a specific technology adoption curve, disproportionately affects the demographic with the greatest usage and the most neurological vulnerability, and has been privately acknowledged by the companies responsible.
The precautionary panics about television were based on speculation. The concerns about social media are based on evidence accumulated over a decade that the platforms have spent considerable resources trying to confuse, dispute, and delay.
What a ban for under-sixteens would not do is fix everything, or even most things. Teenagers over sixteen would still be exposed to the same algorithmic architecture. The social comparison economy would not disappear. The data harvesting would continue.
The advertising model would remain intact. A minimum age restriction is not a complete solution to the harms of attention-economy social media. But it is a meaningful, enforceable starting point that establishes the foundational principle that children are not a legitimate commercial target for addiction-engineered products. It is the step that precedes all the other steps. And it is the step that the platforms, with their billions in lobbying budgets and their armies of communication professionals, will resist with every available instrument until the political cost of inaction exceeds the political cost of acting.
That calculus is changing. Australia changed it. The United Kingdom is changing it. The question for every other government is not whether the evidence supports action. The evidence has supported action for years.
The question is how many more children have to visit emergency departments before the arithmetic of political will finally shifts.
There is a child, right now, in a bedroom somewhere, in the blue light of a phone that knows more about her psychological vulnerabilities than her parents do, being served content by an algorithm that has been optimised over billions of iterations to find exactly the image, the video, the comment thread, that will keep her awake for one more hour.
The platform running that algorithm earned revenue from the advertising displayed alongside that content. The company that built the algorithm knew, from its own research, that content of this type made girls feel worse about themselves.
Nobody in a position of regulatory authority required them to stop. We know all of this. The only remaining question is what we intend to do about it.
0 Comments