The Policy at a Glance
On June 15, 2026, Prime Minister Keir Starmer and Technology Secretary Liz Kendall announced that the UK will prohibit children under 16 from using Snapchat, TikTok, YouTube, Instagram, Facebook, and X. Private messaging services WhatsApp and Signal are exempted. The announcement goes beyond an access block: it prohibits livestreaming for under-16s, mandates default-restricted settings for 16- and 17-year-olds on intimate platform features, blocks stranger-initiated contact on social and gaming platforms, and imposes an 18-year floor on AI 'romantic companion' chatbots.
Legislation goes to Parliament before Christmas 2026, with protections taking effect in Spring 2027. Ofcom, the UK's communications regulator, will enforce compliance under the existing Online Safety Act framework. Platforms that repeatedly fail to exclude under-16s face substantial fines, and in the most egregious cases Ofcom can seek judicial orders to block a platform from the UK market entirely. Ofcom has also been tasked with publishing an urgent review of 'effective age assurance' options — to be delivered in the coming months.
The Strongest Case for Acting
The political mandate is unusually broad. A national consultation drew more than 116,000 responses. Nine in ten parents back the ban, and two-thirds of young people themselves support restrictions on under-16 access. Consensus at that scale is rare for any tech regulation, and the government is right to take it seriously.
The most persuasive argument for a hard platform ban — rather than weaker alternatives — is that incremental measures have not worked. The Online Safety Act 2023 imposed substantial child-safety duties on platforms. Ofcom's enforcement has been slow and complaint-led. Parental controls and screen-time dashboards have not meaningfully altered structural exposure to algorithmic amplification. By placing legal liability squarely on platforms, the new regime aligns incentives in a way no prior measure has managed. The government's instinct to put responsibility on companies rather than on parents is the right one.
What Australia's Model Actually Showed
The UK explicitly models its approach on Australia's, which enacted the world's first enforceable under-16 social media ban in December 2025. According to Australia's eSafety Commissioner, platforms removed access to 4.7 million under-16 accounts in the opening weeks of enforcement. Platforms face fines of up to A$49.5 million for failing to take 'reasonable steps' to prevent under-16 access — a standard that requires real effort without demanding perfection.
The detail matters. 'Reasonable steps' is deliberately flexible: it does not mandate a single verification mechanism or guarantee clean exclusion. In practice, platforms deployed a patchwork: facial-age estimation via selfie, linked bank-card data, account-age inference, and uploaded government IDs. Snapchat, for instance, routes users through three separate verification pathways using third-party provider k-ID. The result is not a clean ban but a friction increase — and documented VPN circumvention by Australian teenagers has been widely reported since day one.
The UK appears to have partly absorbed this lesson. The June 15 announcement expressly defers the mechanism question to Ofcom, acknowledging the government does not yet know what 'effective age assurance' will mean in practice.
The Unresolved Mechanism Problem
This deferral is where the policy is most exposed — and where the criticism is most legitimate.
The Electronic Frontier Foundation observed that there is currently 'no reliable, privacy-preserving method' of verifying age online at scale. Every mechanism that actually works — document upload, biometric estimation, financial account linkage — requires collecting data on adult users who are not the target of the regulation. The UK's own fact sheet acknowledges this implicitly: verification 'could be as simple as a facial recognition check for over-18s' or rely on a linked credit card. Both options are data-intensive by design, applied to tens of millions of adults to filter out a minority of underage users. That creates aggregated datasets with significant breach and misuse risk that Parliament must scrutinise before regulations are laid.
The EFF also documents 'significant scientific doubt' about the causal link between social media use and adolescent mental health decline — the empirical premise underpinning the entire policy. One critical literature review, cited in the EFF's analysis, described the policymaking trajectory as 'panic first, evidence later.' That is not a case for inaction — children do face real algorithmic harms, and platforms have repeatedly exploited adolescent psychology for engagement. It is, however, a strong argument for building rigorous pre-specified evaluation into the regulations from the outset, with published outcome metrics at 18 months. No such provision has been announced.
Scope and Substitution Risk
The announced ban names six platforms. Discord, Reddit, Roblox, Twitch, Kick, and the broader landscape of gaming community platforms — where substantial adolescent social activity already occurs — are not captured by the June announcement, even though the Online Safety Act imposes broader duties on user-to-user services generally.
A list defined by today's market leaders will not map onto where under-16s actually spend their time in 2027. Teenagers displaced from Instagram are not necessarily moving to safer environments — they may migrate to less-regulated spaces. A policy scoped by platform characteristics (algorithm-driven public feeds, unlimited stranger contact, social graph formation) would be more durable than one scoped by brand name.
What Proportionate Regulation Looks Like
The government's intent is well-founded, and the political mandate is real. But the June 2026 announcement creates obligations without mechanisms, sets a Spring 2027 deadline for a regulatory architecture Ofcom has not yet designed, and restricts access to distribution channels without imposing design obligations on the algorithmic systems that actually drive harm. Banning under-16s from the front gate while leaving the attention-optimisation engine intact treats the symptom rather than the cause.
A more durable approach would add three elements the current plan lacks: mandatory, privacy-specified age-assurance standards — defined in the legislation rather than deferred — so that Parliament, not Ofcom alone, sets the privacy ceiling; algorithmic-design obligations for covered platforms, including chronological feed options and hard caps on engagement-optimised recommendations to minors; and a statutory sunset review with published outcome metrics. The UK public wants action. Spring 2027 needs to deliver a defined mechanism that protects children without requiring adults to pass an identity check to use the internet.