Prime Minister Keir Starmer's package carves out messaging apps and AI chatbots differently, exposing the government's theory of which product designs pose the actual risk.
The UK government on Monday announced a ban on social media use for children under 16, applying to Snapchat, TikTok, YouTube, Instagram, Facebook, and X, while leaving messaging services such as WhatsApp and Signal untouched and restricting AI "romantic companion" chatbots to users 18 and over (TechCrunch; UK government press release). The carve-outs are not arbitrary. They expose the government's working theory of which product designs actually drive the harm it is trying to stop.
Messaging apps, which the policy exempts, are end-to-end-encrypted communication tools rather than algorithmic public feeds. The AI "romantic companion" rule, by contrast, applies a stricter age threshold to a category that did not exist in earlier regulatory debates, and treats synthetic conversational companions as continuous-attention products in their own right. Read together, the platform list, the messaging exemption, and the chatbot restriction each signal where the government thinks the harm mechanism lives in the design of the product, not just in the content it carries.
The government's stated rationale is parental authority and childhood itself. "Put power back in parents' hands and give kids the childhood they deserve," Starmer said on Monday, framing the policy as a transfer of control from platforms to families. The implementation target is potentially in place by Spring 2027, per the government, enacted through secondary legislation under the Children's Wellbeing and Schools Act (UK government press release).
Whether the package can deliver on that framing is the central test, and the one critics say will determine whether it works in practice. Three specific questions will define the outcome.
The first is age assurance. The government says it will introduce "highly effective age assurance (HEAA) measures" and has asked Ofcom to conduct a rapid study on what constitutes effective age verification. A ban enforced only at account creation, rather than at the session level, will leak. The government has said it will ensure Ofcom has the funding it needs for enforcement.
The second is platform compliance. The named services operate globally, and each will need to apply UK-specific age-gating to its UK user base. The government is examining the Australian model as it determines which platforms fall under the restrictions. Past attempts at jurisdictional age rules have produced uneven enforcement and frequent workarounds.
The third is migration. Children who cannot access the named platforms in the UK can route around the block through virtual private networks, alternative front-ends, or simply by using a parent's or older sibling's verified account. Critics of age-based online safety rules have argued for years that the demand for these services does not disappear when the legal access disappears, and that the worst outcomes from a ban can land on the children who do not have a compliant household.
The policy also goes beyond a blanket social media ban with additional restrictions on livestreaming and stranger communication for under-16s, and on a wider range of online services including gaming sites. Restrictions on these functionalities will also apply by default to 16- and 17-year-olds to prevent a cliff-edge at 16. The government said it is examining overnight curfews and breaks in infinite scrolling for under-18s, with more detail to come in July.
The UK is the second country to attempt an under-16 social media ban, following Australia, which imposed the first such restriction late last year. Canada, France, and Denmark are developing their own versions. The UK government says its package goes further than any of those — and that it will learn from Australia's experience in implementing HEAA. Whether that breadth survives contact with the enforcement problem is the measure by which the policy will be judged.