The Enforcement Problem No One Wants to Name
What happened
In the same week, the UK government committed to imposing age or functionality restrictions on social media for under-16s regardless of its ongoing consultation, Norway announced plans to introduce under-16 legislation requiring tech companies to enforce age verification, and Manitoba Premier Wab Kinew announced Canada's first provincial ban on social media and AI chatbots for minors. All three announcements followed Australia's December 2025 Social Media Minimum Age law, the first of its kind globally. An Australian government report published this month found that approximately 70% of Australian children who had social media accounts before the ban still have access, and some platforms have been encouraging children to retry verification until they are approved. A 2025 hack of a UK and Australia age-verification firm exposed the government IDs of 70,000 Discord users.
Every government announcing a youth social media ban this week already knows bans do not work, because Australia ran the experiment first. The political decision to announce bans is not driven by evidence of effectiveness. It is driven by parental anxiety, tragic case visibility, and a calculation that announcing action costs less than defending inaction, even when the action fails.
The Hidden Bet
Age verification technology will improve enough to make bans enforceable
Age verification at scale requires either government ID upload, facial estimation, or third-party services. All three options have significant privacy and accuracy problems. Facial estimation is notoriously unreliable at distinguishing 15-year-olds from 16-year-olds. Third-party services are already getting hacked. There is no technical trajectory that makes this problem easy.
Banning platforms for under-16s reduces social media harms to that age group
Australia's eSafety Commissioner found no observable change in cyberbullying or image-based abuse complaints in the months following the ban. VPN downloads tripled as children circumvented it. Children who circumvent bans do so without the parental oversight the ban was designed to restore. Circumvention may produce worse outcomes than access.
Social media companies will enforce bans more effectively than they have historically enforced age policies
Meta, TikTok and others have had minimum age requirements for years. They have not been enforced at meaningful scale. The incentive to grow monthly active users conflicts with genuine age enforcement. Fines are small relative to revenue. Some platforms, according to Australian reports, are actively undermining their own verification systems.
The Real Disagreement
The fundamental fork is between two interventions that are not equally politically available. Evidence-based approaches to reducing social media harm in children include algorithmic curation restrictions, default privacy settings, parental control tools, and time limits, all of which already exist in some form and none of which require age verification infrastructure. Bans are politically more visible and feel more decisive, but the evidence says they do not work. The real disagreement is not about whether children should be protected from social media harm. It is about whether the political benefit of announcing a ban justifies imposing a surveillance infrastructure on all adults who must prove they are not minors to access the internet. The lean here is that the ban approach is the wrong intervention chosen for the right reason, and the people most harmed by its failure are the children it was meant to help.
What No One Is Saying
Every government that announces a social media ban is effectively outsourcing child protection to the platforms it is trying to regulate. Age verification requires trusting the companies whose design choices caused the problem in the first place to police their own user base honestly. The governments that are serious about this problem should be funding independent age-verification infrastructure, which the EU is moving toward, rather than issuing mandates and hoping platforms comply.
Who Pays
All adult internet users
When legislation takes effect; UK consultation deadline is May 26
Age verification systems that verify children are not present require adults to prove they are not children. This means uploading government identification, biometric data, or third-party verification to access social media and, in some proposals, messaging apps and gaming platforms.
Children who circumvent bans without parental knowledge
Immediately after any ban takes effect
Children who use VPNs to access banned platforms do so outside normal app stores and parental visibility. They may access more harmful content and less moderated environments than the mainstream platforms the ban targeted.
Age verification services
Ongoing, with breach risk growing as mandates spread
Centralized repositories of government IDs and biometric data for millions of users create high-value hacking targets. The 2025 breach that exposed 70,000 Discord users' government IDs is an early indicator of the systemic risk this infrastructure creates.
Scenarios
EU verification passkey becomes global standard
The EU's independent age-verification system, which works like a passkey without requiring platforms to hold ID data, proves effective in early deployment. UK, Norway, and Manitoba align their systems with it. The debate shifts from bans to verified-access infrastructure.
Signal EU announces measurable enforcement success rates above 85% in the first six months of the passkey system.
Bans fail visibly, political backlash builds
UK and Norwegian bans go into effect and produce the same 70% circumvention rate Australia found. Parliamentary scrutiny intensifies. A data breach at a UK age-verification provider exposes millions of adults' IDs and triggers a class-action lawsuit.
Signal UK eSafety Commissioner publishes a compliance report showing no material reduction in under-16 platform access six months post-implementation.
US federal action on algorithmic restrictions instead
The bipartisan US federal legislation focuses on algorithmic recommendation bans and privacy protections rather than age-based bans, following the Stars and Stripes / First Amendment analysis. The EU passkey model and the US algorithmic model diverge, creating two competing global frameworks.
Signal The US federal bill passes committee with the age-ban provision removed and algorithmic restriction provisions strengthened.
What Would Change This
If controlled studies in Australia or another jurisdiction demonstrated that bans reduced rates of depression, anxiety, or online abuse in the affected age group compared to control groups, the case for bans as effective policy would strengthen materially. The current evidence shows no such effect.
Related
The Ban That Moves Children Underground
decisionThe World Is Copying Australia's Teen Social Media Ban. Australia's Is Already Failing.
decisionParliament Voted Down the Ban. Starmer Called Tech CEOs Into Downing Street Anyway.
powerNorway, Indonesia, and Australia Are Banning Social Media for Under-16s. Meta Is Threatening to Leave States That Try It.