Social media bans: Protecting children or building a panopticon

By Dr Sanjana Hattotuwa Dr Sanjana Hattotuwa Australia’s world-first under-16 social media ban has produced a statistic Sri Lankan legislators, and our Prime Minister, should carefully consider with before expressing their interest in implementing a similar law. A peer-reviewed…

By Dr Sanjana Hattotuwa Dr Sanjana Hattotuwa

Australia’s world-first under-16 social media ban has produced a statistic Sri Lankan legislators, and our Prime Minister, should carefully consider with before expressing their interest in implementing a similar law. A peer-reviewed BMJ study published in June followed 408 adolescents before and three months after the law took effect. More than 85% of those under 16 still used the platforms the ban covers, most through their own accounts; two thirds met age verification from social media platforms no more demanding than declaring a birthday. If a state far better resourced than ours cannot enforce a law such as this, any claim that Sri Lanka will is not a serious one. The study had a modest sample, and captured early effects, and with those commenting on it suggesting governments, including Australia’s own, may need a decade or more to know whether the law shifts the norms of a generation. While this may be the case, a generational hypothesis cannot show that today’s restrictions work, nor answer what children lose, where they migrate, or what surveillance infrastructure the state builds while it waits. An instructive contrast comes from the EU. Receiving a report from an expert panel of doctors, academics, youth representatives and parents she commissioned, the President of the European Commission Ursula von der Leyen called for “phased and gradual access” to social media – supervised use for children aged 3 to 12, “evolving autonomous use” for adolescents to 18 on platforms with meaningful safety features. The panel pointedly declined to recommend a blanket ban; von der Leyen declined to back one, despite pressure from France and Greece for an Australian-style prohibition across the bloc. Policymakers in our country should study the sequence: evidence first, obligations on platforms over prohibition, legislation after summer this year. Australia has already legislated, France and Denmark are deliberating, and the EU is drafting. What Sri Lanka seems keen to import is the prohibition alone, shorn of the expert panels, evidence reviews and public deliberation that preceded it everywhere else. I put a simpler question to BBC Sinhala when discussing MP Faizer Mustapha’s private member’s bill to ban social media for under-16s here. How, precisely, would it work? The conversation in Sri Lanka, as in many countries, has skipped this part. The uncomfortable answer is that only one mechanism exists. To keep under-16s off a platform, you must know the age of every person on it – not just the children, but all of us, through age-gating that obliges the entire user base to prove who they are, ordinarily an NIC, driving licence or passport, submitted to the platform or some intermediary. There is no version of this that checks only the young. Verifying the many is the price of excluding the few. And therein lies the rub. In a country whose democratic institutions have been enervated by emergency powers, surveillance and selective coercion, a mechanism that binds legal identity to online presence is an instrument, not a neutral technical fix. Once the state can attach a real name to an account, it can – should it choose, and quite easily – restrict precisely the people it wishes to silence. Given the enduring post-war militarisation, and rampant surveillance over decades, what stops an age-assurance system in Sri Lanka becoming an identity layer for policing speech, profiling dissent or restricting particular communities? Governments elsewhere also offer no straight answer to that charge, and challenge. Not unlike the draconian Online Safety Act (OSA), and the manner it was rushed through, child safety can be a politically attractive label for an architecture of Orwellian control. Many advocates reach for tobacco to defend these bans, whichhowever is an analogy that obscures more than it explains. Cigarettes do not give a rural teenager a learning community, connect a queer child to support unavailable at home, help a disabled young person overcome physical isolation, or carry political information to someone approaching voting age, and may already be politically vocal or active. Social media does harm, but it also forms part of the social and civic infrastructure young people inhabit, and 16-year-olds today have grown up with. Treating it as a single toxic substance produces correspondingly crude policy. The harm is real, and linked to what I’ve studied on social media since 2012 that, with AI, has got progressively worse. Academic research describes, amongst other deleterious effects, exposure to self-harm content, explicit sexual material, violence, bullying, sleep disruption, compulsive use,and behavioural profiling. Other research ties harmful exposure to personalised feeds, autoplay, infinite scroll and recommender systems – and it is design, not children, that the EU has moved against. Days before receiving its expert panel’s report, the European Commission warned Meta that the “addictive” architecture of Facebook and Instagram – infinite scroll, autoplaying video, recommendations that shift the brain “into autopilot mode” – encourages compulsive use amongst younger users, and must change or attract fines of up to 6% of global annual turnover. Meta, predictably, disagrees. Mental-health studies record correlations between heavier use and anxiety or depression without settling on a single causal story. “Social media” spans algorithmic feeds, private messaging, livestreaming and mental-health support; its harms follow different pathways. A law that bundles them into one problem, with one age threshold as the cure, does not, and cannot demonstrate necessity or proportionality. There’s also something called children’s rights which no policymaker interested in policing social media ever talks about. Article 12 of the Convention on the Rights of the Child requires states to hear children in matters that affect them, and General Comment No. 25 extends that obligation to the digital environment. Adults and children often understand risks, benefits and remedies differently: young people support protection, yet many prefer control over algorithms, better reporting and school support to government control of their online lives. Consultation with youth must shape law-making, and not just occur performatively after an awful law has been hurriedly passed. Sri Lanka has better options, with the EU highlighting the principle – platforms “must prove that their services do no harm”. Parliament or a body of policymakers, institutions like the Child Protection Authority, and leading civil society researchers working in concert could require child-appropriate defaults, private accounts, limits on contact from unknown adults, rapid complaint and redress, restrictions on behavioural advertising and independent audits of recommender systems. Joining many other independent regulators, Sri Lanka could ask platforms to constrain autoplay, infinite scroll and the repeated recommendation of self-harm or sexual material. These measures target the mechanism of harm. Where a specific risk justifies age assurance, we can demand privacy-preserving proof of an age range, not identity transfer – data minimised, reuse forbidden, deletion enforced – and prohibit any central database connecting online accounts to state identity. In simplistically focussing solely on the digital or online harms, policymakers in Sri Lanka never seem to even remotely appreciate how regulation alone doesn’t meet needs schools, families and public services neglect: education on privacy, algorithms, synthetic media, consent and harassment; support for parents, teachers and counsellors; investment in mental-health care, good public libraries, sport, and safe public spaces. Under-16s, and youth in Sri Lanka more generally, are dealing with what earlier generations – on the whole – did not: the aftershocks of 2022’s economic collapse, the rapid rise of AI, climate change, and a Gordian knot of socio-economic, political and existential crises that manifest as mental-health issues and stress no app can be blamed for in full. Heavy use may signal isolation, the absence of credible alternatives, or a generation self-medicating against circumstances adults created. A social media ban moves the symptom elsewhere while leaving the causes, and driversuntouched. To this end, what should animate policymaking, I stressed to BBC Sinhala, is grounded evidence – and in Sri Lanka there is a clear vacuum. MP Mustapha or the Prime Minister have not, to my knowledge, produced any domestic evidence base establishing that a ban will address what they say it will. No settled consensus holds that removing minors from these platforms resolves the harms attributed to them; the debate is live, and thoughtful people disagree. The platforms themselves possess no reliable method to do what such a law demands. The countries that legislated did so after formal study. Sri Lanka has commissioned none. To legislate on the strength of personal conviction, before research establishes that the problem exists in the form imagined, is to get the sequence exactly backwards. Study first to establish the precise nature of the problem. Legislate, if at all, afterwards. If the government still wants a general ban, it should publish the evidence threshold the law must meet, conduct children’s-rights and data-protection impact assessments, run a transparent pilot with a sunset clause, and fund a national study reaching beyond Colombo across provinces, languages, incomes and family circumstances. If legislators cannot do this, they do not have a child-protection framework. They have a ban, an identity system to gate-keep what millions of citizens can access online, and a familiar invitation to trust political power without evidence, which is protection as pretext, prohibition as performance, and a panopticon in production. Sri Lanka has tried this before, many times over, and no one in society is better off for it.