A Softer Rule for the Age Group Just Above the Ban
On 15 July 2026, the Department for Science, Innovation and Technology announced that 16- and 17-year-olds in the UK will get a default midnight-to-6am lockout from social media, alongside autoplay and personalised infinite feeds switched off by default, under what officials call the "transition tier" leading into the under-16 social media ban announced by then-Prime Minister Keir Starmer on 15 June 2026. The first regulations are due before Parliament by the end of 2026, with both measures entering force in spring 2027, using powers under the Children's Wellbeing and Schools Act rather than a new statute (per the gov.uk curfew announcement).
Crucially, the curfew is a default, not a prohibition. Teenagers aged 16 and 17 can switch it off in their account settings. Technology minister Kanishka Narayan framed this explicitly as an empowerment measure rather than a lockout, citing pilot data showing that more than 90% of teenagers kept the restrictive default settings active once enrolled. That distinction — opt-out defaults versus a hard ban — is the single most important design choice in this announcement, and it deserves credit before any criticism follows.
The Case For It, Stated Fairly
The government isn't inventing a problem. A pilot of more than 300 teenagers and parents found that the overnight curfew produced the clearest, most consistent sleep-related improvements of any intervention tested — participants reported falling asleep faster, waking more easily, and feeling less tired the next day, according to the gov.uk release. Dame Rachel de Souza, the Children's Commissioner for England, called the announcement "a positive step" while reserving judgment on delivery — a measured response from an official who has previously pushed for stronger platform accountability. And the design logic — avoiding a "cliff edge" where protections that applied at 15 vanish overnight at 16 — is coherent. Sleep-hygiene interventions with opt-out defaults are a substantially lower-friction policy tool than the blanket under-16 ban this curfew sits alongside, and the behavioral-economics case for defaults (most people keep whatever setting is pre-selected) is well established, not speculative.
Where the Design Still Strains
The catch is that a default only works if the platform can reliably identify who is 16 or 17 in the first place — and that requires the same age-assurance infrastructure the under-16 ban depends on. Ofcom itself, in its statement responding to the government's announcement, said current systems are unlikely to be "highly effective" enough to support a meaningful age threshold at the point of entry, and the Secretary of State has asked Ofcom for a rapid study on stronger age-assurance methods by October 2026 — a study commissioned after the policy was announced, not before.
That sequencing matters. Open Rights Group has argued that expanding age-gating to distinguish 16-and-17-year-olds from adults necessarily means collecting more identity data from everyone nearby that boundary — a point the group has made in the context of the wider ban, but which applies with equal force here, since a curfew that only some users see requires the same verification layer as a ban that blocks some users outright. The group's warning that "we already know these systems are risky," pointing to a Discord age-verification breach, is not a hypothetical concern; it's a description of the exact infrastructure this curfew needs to function as anything other than a checkbox in a settings menu.
"Protecting children online should not mean building surveillance infrastructure for everyone." — Open Rights Group
There is also a harder practical question the government has not yet answered in public: what happens when the default is trivially reversible by anyone who can find the settings menu, including under-16s who have already evaded the primary ban's own age checks? A 90% retention rate in a pilot with engaged, consenting families is not the same as a 90% retention rate at national scale, where the curfew's actual behavior-shaping power depends entirely on how quietly the opt-out is buried — which is itself a dark-pattern-adjacent design choice regulators claim to oppose elsewhere in the same policy package.
The Better Version of This Policy
None of this argues for scrapping the curfew. Opt-out sleep defaults for older teens are a genuinely proportionate intervention, and the government deserves credit for choosing nudge over ban at this age tier. But Parliament should insist that Ofcom's age-assurance findings — due in October, months after this announcement — inform the regulations laid before the end of 2026, not the other way around. A default that works depends on verification infrastructure that doesn't yet meet Ofcom's own bar for reliability. Legislating the policy before the verification method is settled risks locking in either a toothless default (easily bypassed) or a heavier identity-verification regime than ministers have been willing to admit to the public. Sequence matters as much as substance.