Modeled on Australia's under 16 ban, the package also covers livestreaming, disappearing messages, and AI companion chatbots marketed to children as ongoing companions, with some restrictions applying to 16 and 17 year olds.
PM Keir Starmer set out legislation on Monday to bar under-16s from social media in the UK, with the rules expected to take effect in spring 2027, according to a UK government press release reported by The Register. The bill is to be introduced before Parliament's Christmas recess. The under-16 ban is the most attention-grabbing piece of a wider package that also targets livestreaming, disappearing messages, stranger contact, and AI companion chatbots.
The plan follows the model of Australia's under-16 social media ban, focusing on user-to-user platforms whose primary purpose is social interaction and user-generated content. Named in scope: Snapchat, TikTok, YouTube, Instagram, Facebook, and X. WhatsApp and Signal are explicitly excluded as messaging services, the government said.
For 16- and 17-year-olds, the same functionality restrictions apply by default so there is no cliff edge at 16. The government is also examining overnight curfews and mandatory breaks in infinite scrolling for under-18s, with more detail expected in July, The Register reported.
The government cites its 'Growing Up in an Online World' consultation as evidence: 91% of parent respondents backed a minimum age of 16, more than four in five said social media's risks outweigh its benefits for children, 88% said restrictions would cut exposure to inappropriate content, 75% said there would be fewer arguments at home, and 77% said schools would find digital behavior easier to manage. Those figures come from a consultation run by the commissioning party rather than independent polling, and are best read as the government's own evidence base.
Enforcement will rest on 'highly effective age assurance' rules set by Ofcom, the UK's communications regulator. An Ofcom spokesperson said the regulator is ready to support the new framework; that is signaling, not a formal decision.
Starmer framed the move as a 'line in the sand' against platform self-policing, accusing social media companies of building products they know harm children. The criticism sits in the wider package: livestreaming, disappearing messages, and AI companion chatbots, services marketed to children as ongoing companions in the Replika or Character.AI mold, are all in scope because the government has decided the harm is in the design, not just the content.
Skeptics quoted by The Register were quick to flag gaps. The Open Rights Group's James Baker warned of surveillance and censorship creep and noted that the business models driving the harms are untouched. Mark Jones, a partner at Payne Hicks Beach, questioned enforceability, pointing to Australia's enforcement problems and to the kinds of low-effort bypasses already documented in UK age-verification trials.
The two tests that will determine whether the package works are age assurance and the AI companion rules. Australia has struggled to keep determined children off banned platforms, and the UK has yet to publish the technical standard platforms will have to meet. The AI companion piece is more open still: it is a category that barely existed when Australia's law was drafted, and the UK's 18+ threshold for romantic-companion chatbots is a policy choice, not a technical one.
Legislation is to be introduced before Christmas, with rules expected in spring 2027. The detail that will decide whether this becomes a genuine shift or another consultation exercise that fades in implementation is what Ofcom puts in the age-assurance standard, not the headline ban.