From Canberra to Sacramento, legislators are racing to erect digital age gates. The policies are popular, the headlines are bold, but the technical reality remains a mess.
Australia’s Senate passed its world-first ban on social media for under-16s in late November with a 34-19 vote. The legislation, which forces platforms like TikTok, Snapchat, and Instagram to take “reasonable steps” to verify age or face fines up to AUD 49.5 million (£25 million), is the most aggressive move yet in a global regulatory wave. Yet, as the ink dries on the statute books, a rare consensus is forming among technologists, human rights advocates, and even some child safety experts: these laws are largely unenforceable, potentially dangerous, and born of political expediency rather than evidence.
The Anatomy of a Moral Panic
The momentum is undeniable. In the United States, Florida has banned accounts for under-14s and required parental consent for 14- and 15-year-olds. Utah and Arkansas have pushed similar verification mandates, though court injunctions have stalled several on First Amendment grounds. The UK’s Online Safety Act, while not an outright ban, imposes a duty of care that effectively pressures platforms to exclude children or face ruinous fines. France has mandated parental consent for under-15s. Norway is proposing a 15-year age floor.
The political logic is straightforward. Parents are anxious. Teachers report attention spans fractured by algorithmic feeds. A steady stream of research — some robust, some contested — links heavy social media use to rising rates of anxiety, depression, and body dysmorphia among adolescents. When the US Surgeon General, Vivek Murthy, issued an advisory in 2023 calling for warning labels on platforms, he handed lawmakers a medical imprimatur for intervention.
“It is a classic moral panic dynamic,” says Dr. Pete Etchells, a professor of psychology and science communication at Bath Spa University and author of *Unlocked: The Real Science of Screen Time*. “We have a genuine problem — youth mental health is declining — and we have a highly visible new technology. The temptation to draw a straight causal line between the two is immense, but the data simply doesn’t support a monocausal explanation.”
The Verification Paradox
The central flaw in every current proposal is the “how.” How does a platform verify a user is 16 without compromising the privacy of the adult population?
Current options form a grim menu. Credit card checks exclude the unbanked and are trivial to borrow. Facial age estimation — championed by firms like Yoti — claims high accuracy rates but struggles with diverse skin tones and facial differences, raising bias concerns. Government-issued digital ID wallets, the preferred route for the EU’s eIDAS 2.0 framework, do not yet exist at scale. Uploading a passport creates a honeypot of sensitive identity documents ripe for breach.
“If you build a system that requires every adult to prove their age to access legal speech, you have built a surveillance infrastructure,” argues Dr. Joss Wright, a senior research fellow at the Oxford Internet Institute. “You are solving a child safety problem by dismantling adult anonymity. That is a disproportionate trade-off.”
The Australian law explicitly bans the use of government ID for verification, a nod to privacy fears. But it leaves the “reasonable steps” definition to the eSafety Commissioner and industry codes. Critics say this kicks the can down the road. “It is regulation by vibes,” says a senior policy analyst at a major platform, speaking on background. “We are being told to wave a magic wand. If we fail, we get fined. If we succeed via invasive tracking, we get sued by privacy regulators.”
The Migration to Darker Corners
There is a second-order effect that receives less airtime: displacement. When mainstream platforms erect hard barriers, teenagers do not log off. They migrate.
Evidence from Utah’s short-lived verification law suggested teens simply used VPNs, borrowed credentials, or flocked to unmoderated alternatives — Discord servers, fringe forums, encrypted messaging apps like Telegram — where safeguards are non-existent and predators operate with impunity. The “safety by exclusion” model assumes a binary world: safe mainstream platforms and a void. The reality is a hydra.
“We risk creating a two-tier internet,” says Beeban Kidron, the crossbench peer who architected the UK’s Age Appropriate Design Code (the Children’s Code). “One tier is a sanitised, verified, data-harvesting walled garden for the compliant. The other is a lawless wild west for everyone else — including the most vulnerable kids who lack the digital literacy or parental support to navigate the mainstream.”
Kidron’s code, which came into full force in 2021, took a different tack: it mandated privacy by design, turned off geolocation by default, and banned nudge techniques for under-18s. It did not ban children. Early evaluations suggest it forced meaningful product changes — Instagram introduced “Take a Break” reminders; TikTok disabled direct messages for under-16s — without the nuclear option of exclusion.
The Algorithmic Elephant in the Room
Beneath the age-gate debate lies the uncomfortable truth politicians are reluctant to regulate: the business model.
Social media platforms are optimised for engagement. The recommendation engines that serve infinite scroll, the variable-ratio reinforcement of likes, the autoplay loops — these are not bugs. They are features designed to maximise time-on-device. For a developing prefrontal cortex, they are potent.
“If you want to protect children, you regulate the architecture, not the demographic,” says Guillaume Chaslot, a former Google engineer who worked on YouTube’s recommendation algorithm and now runs AlgoTransparency. “Ban the infinite scroll for minors. Ban autoplay. Mandate chronological feeds. Force transparency on what the algorithm optimises for. That is technically trivial compared to age verification. But it hits revenue.”
The EU’s Digital Services Act (DSA) gestures toward this, requiring “very large online platforms” to offer a non-algorithmic feed option and conduct systemic risk assessments. The UK’s Online Safety Act includes similar duties. But enforcement is nascent, and the platforms are litigating hard.
In the US, the Kids Online Safety Act (KOSA) has passed the Senate with overwhelming bipartisan support (91-3) but stalled in the House amid lobbying from tech groups and civil liberties organisations who fear its “duty of care” provision could be weaponised to censor LGBTQ+ content or reproductive health information. The debate illustrates the minefield: a law designed to protect children can become a tool for political censorship if drafted loosely.
The Evidence Gap
Perhaps the most striking aspect of the current wave is the paucity of rigorous evidence supporting bans.
A 2024 review by the UK’s Chief Medical Officer found “insufficient evidence” to support a blanket ban on phones in schools, let alone a societal ban on social media. The American Psychological Association’s 2023 health advisory noted that social media is “not inherently beneficial or harmful” — its effects depend on content, features, and the individual child. Longitudinal studies like the ABCD Study in the US show small effect sizes, often drowned out by socioeconomic factors, family dynamics, and academic pressure.
“Legislators are treating social media like lead paint or tobacco,” says Etchells. “But it’s not a toxin with a linear dose-response curve. It’s a complex social environment. Some kids find community; others find bullying. The same platform can be a lifeline for a trans teen in a rural town and a vector for anorexia content for another. A blunt ban helps neither.”
Why it Matters
The rush to ban children from social media reflects a legitimate societal cry for help — but the prescribed cure risks being worse than the disease. By outsourcing parenting to clumsy age gates, governments are normalising universal identity verification, eroding digital privacy for adults, and pushing vulnerable teens into unregulated shadows, all without credible evidence that exclusion improves mental health. The real regulatory fight — forcing platforms to redesign addictive architectures, audit opaque algorithms, and prioritise safety over engagement — is harder, slower, and less headline-friendly. But it is the only path that protects children without breaking the internet for everyone else.
