The UK government will ban social media for children under 16, but that headline understates the policy. The package announced by Prime Minister Keir Starmer on Monday is a three-tier architecture that also restricts under-17s from live-streaming and chatting with strangers online, and bars under-18s from using AI chatbots marketed as romantic or sexual companions. Each tier reflects a separate policy judgment about which digital experiences the state is willing to restrict, and the layering makes the UK regime more granular than Australia's late-2025 under-16 law.
Legislation to enforce the restrictions will be introduced to Parliament before Christmas, with the protections in force by spring 2027, according to the UK government's announcement.
The first tier targets the platforms most British teenagers actually use: Instagram, Facebook, YouTube, TikTok, Snap, and X. Messaging services such as WhatsApp and Signal are explicitly excluded, a carve-out that signals where the government believes surveillance and privacy trade-offs would outweigh the protective aim.
The second tier, covering under-17s, blocks live-streaming and direct communication with strangers by default. The rule extends to gaming sites that offer those features, closing a route that has historically let under-16s reach adult contact through gameplay overlays and chat.
The third tier, the one with no clear global precedent, sets a minimum age of 18 for romantic companion AI chatbots, the products marketed as partners, roleplay partners, or simulated lovers, distinct from general-purpose assistants. Officials argue that synthetic intimacy with a chatbot is qualitatively different from other online harms, and that the absence of a real human on the other side of the conversation changes the consent and developmental calculation.
Starmer framed the package as giving children "more time for play, less time for scrolling," according to the CNET report of the announcement. The political cover for the policy is a UK government survey of more than 116,000 people that found roughly 9 in 10 British parents support a ban. The figure is government-commissioned rather than independent polling, a caveat that matters for readers weighing how settled the public mood actually is.
Australia is the reference case. Its under-16 social media law took effect in late 2025, and UK officials have framed it as the experimental testbed whose age-assurance approach the UK will borrow from. As the CNET explainer on the Australian law details, the Australian regime relies on platform-side age verification rather than a national identity check, a model that has been criticized for accuracy and for the data it requires platforms to hold. Whether that approach will work at UK scale is one of the unresolved questions the legislation will have to answer.
Industry reaction has been mixed. Snap has argued the policy will push young users toward less safe platforms. YouTube has warned that a flat block would deny teenagers a curated, supervised experience the company already offers. Meta, TikTok, and X did not respond at the time of reporting, a silence that is itself part of the negotiation. The companies that did respond have all raised the same underlying concern: blanket blocking treats users as a single class, when the evidence base for harm varies sharply by product and by age within the under-18 group.
Civil-society critics have gone further. NSPCC chief executive Chris Sherwood warned that a blanket under-16 ban could isolate LGBTQ+ and neurodiverse teenagers who rely on online communities they cannot easily find offline, and argued that the policy risks hiding abuse rather than surfacing it, in the charity's published response. Amnesty International UK director Kerry Moscogiuri called the move "the right diagnosis, the wrong prescription," contending that the underlying problem is platform design, not youth access, in the organization's statement.
Several enforcement questions remain open. The UK has not said how it will handle virtual private networks, which under-16s can use to route around domestic blocks. The age-assurance technology the policy depends on is improving but still produces false negatives, and the privacy cost of any working system is non-trivial. The companion-AI rule, in particular, will have to grapple with products hosted outside UK jurisdiction, where domestic enforcement has limited reach.
A separate workstream due in July will explore overnight curfews and forced breaks on infinite scrolling for under-18s, an indication that the government views the current announcement as the floor, not the ceiling, of its youth-internet agenda.
The package lands three months after Starmer gave tech firms an ultimatum on preventing nude image sharing among minors, an order layered on the existing Online Safety Act age-verification regime. The progression, from a one-week deadline on a specific harm, to a tiered ban on whole categories of platform use, to a forthcoming curfew regime, suggests the government is moving from product-by-product pressure to structural default-setting. Whether the architecture holds depends on age-assurance accuracy, on outcome data from Australia that does not yet exist, and on whether the third tier, the romantic-companion AI ban, can be enforced against products that do not have a UK office to regulate.