Field Notes

question

Do under-16 social-media bans reduce measured harm?

StatusWorkingUpdatedStands on3 sourcesSupportsnotes

Australia prohibits under-16s from holding accounts on major social platforms, France has adopted an under-15 ban, and the United Kingdom has announced an under-16 ban to take effect in spring 2027, plus a midnight-to-6am curfew and off-by-default autoplay and infinite scroll for 16- and 17-year-olds, a package opened by the perceived gaps of the UK Online Safety Act. The decision facing other governments is whether to follow. That decision should turn on whether a ban changes measured harm, not on whether it polls well or whether blocking feels like action.

Why it matters

Less harmful social media requires a blanket gate to outperform less intrusive interventions after circumvention, displacement, and foreclosed benefits are counted. Until recently that test was hypothetical: no country had run the experiment. Now several are running it, and their outcomes will either discipline or vindicate the vault’s skepticism. The question is distinct from Does age verification reduce harms other measures cannot, which tracks age gates on adult content. A social-media ban targets ordinary communication platforms, and its costs fall on friendship, support, and participation, not only on access to pornography.

What is already known

  • UK children bypass age checks at scale: 32% within two months, mostly by low-tech means, and 46% consider the checks easy to beat.
  • Support and skepticism coexist: 62% of UK parents backed an under-16 ban in late 2025, yet many supporters doubted it would work, citing VPN workarounds and displacement toward harder-to-regulate services.
  • France’s implementation problems were visible on adoption day: identifying existing under-15 accounts and verifying every new account are unresolved, and a constitutional review is pending.
  • Children themselves value feature-level restrictions, with 77% welcoming contact limits and 74% welcoming limits on features such as livestreams, and they describe engagement mechanics as their most immediate day-to-day problem. All four points come from the Internet Matters Online Safety Act report.

What would settle it

  1. Pre- and post-ban measurement in Australia and France of wellbeing, sleep, harassment exposure, and unwanted contact, not only account counts or platform compliance.
  2. Circumvention and displacement measurement: false-age accounts, borrowed credentials, migration to unregulated or fringe services, and VPN use by age group. Displacement sets what this has to observe: a catchment decided in advance — which services count as the destination — because relocation is only detectable somewhere an evaluation is looking, and it needs the opposite spillover counted too.
  3. Benefits foregone: effects on social connection, support-seeking, and participation for the ages excluded.

  4. Whether bans outperform the design-level interventions tracked in Which social-media designs reduce harm for young users on the same outcomes.

  5. Find Australia’s post-ban evaluation data on wellbeing, exposure, and circumvention, not only compliance reporting.

  6. Track French implementation after the constitutional review and the September 2026 school year.
  7. Track the UK implementation after the spring 2027 effective date, including circumvention of the under-16 ban and the curfew for 16- and 17-year-olds.

Built on 3 sources (3 external).

Working out connections…