Defend Digital Me · Briefing

The UK Under-16s Social Media Ban: Questions & Answers

A children’s-rights response to the government’s announcement of 15 June 2026 — what was announced, what the evidence actually shows, what age verification really costs, and why a ban is the wrong tool.

Published 15 June 2026 · For media, parliamentarians, parents and practitioners

On 15 June 2026 the government announced a ban on social media platforms offering their services to under-16s, alongside restrictions on specific functions and a minimum age of 18 for “romantic companion” AI chatbots. We support protecting children online. We do not support this ban. The weight of considered expert evidence points away from blanket prohibition and toward fixing the underlying causes of online harm. This briefing answers the questions we are asked most often.

Questions for journalists to ask

Ministers have announced the policy. These are the questions still unanswered — worth putting to government, Ofcom and on record.

1. Australia’s regulator reports around 70% of parents say their child still has an account six months on. Why impose a ban when it does not keep children off social media? What is the UK’s measurable test of success, and at what point would the government accept the policy does not work?
2. How was it decided which specific platforms are in and out of scope — and how will the government avoid a “whack-a-mole” problem as children migrate to unlisted services?
3. YouGov polling puts UK parental support at 76%, not the 90% being widely reported, and finds parents split 45%–46% on whether a ban would even work. Why lead with the higher, non-representative figure?
4. Why announce this now, when the consultation evidence has not been analysed and the ban in Australia shows it is not meeting any “child safety” goals?
5. Does a ban mean the government has given up being able to regulate and to enforce its own law — that the Online Safety Act, which already carries extensive content restrictions, has failed? Does it hand companies a get-out-of-jail-free card to carry on being toxic, while giving children a more dangerous cliff-edge from nothing to the full adult experience at 16?
6. The announcement appears to bring games into the under-16 ban — minimum age limits, plus a rule that children can only play with people they already know. What does that mean for companies now needing to know about the relationships between players?
7. To verify a user is over 16, every adult must also prove they are not a child. What is the privacy impact for the whole population, and has the government carried out an assessment of it?
8. Ofcom must report on the effectiveness of age assurance by 17 July 2026. Why legislate before that statutory evidence is public, and how will it feed into the policy impact assessment?
9. Who owns the age-check process for each of the ten in-scope sites, and what is their business model?
10. What is the appeal route when age checks make mistakes — in either direction?
11. Children’s biometric data and identity documents are high-value and high-risk: they last a lifetime and, once lost, are lost for life. Will the government ban biometric processing for social media age checks, if children’s safety is a genuine priority?
12. Who will children sue for the loss of their identity data in the event of a breach, as a child or as a future adult — the age verification company or the social media platform?
13. Were children consulted and their views given due weight before the decision, as Article 12 of the UNCRC requires, given a minister said before the consultation closed that a restriction would be imposed “under any outcome” — when children asked for feature restrictions, not bans?
14. How has the government weighed the evidence that excluded children will lose access to news and community — when Australian research shows one in two affected children now see less news?
15. Three of the UK’s four Children’s Commissioners (Scotland, Northern Ireland and Wales) oppose a blanket ban. Why is the government citing the only one Commissioner who supports it, while three who engaged children’s own evidence oppose it?
16. The Minister for AI and Online Safety has claimed on Newsnight (15/6/2026) that age verification is about, “checking ages so young people are kept from child abuse”— which sounds laudable until you ask why, where, if such content is illegal and should never be seen, shared or accessible by anyone?

The announcement & the evidence

What did the government actually announce?

A ban on social media platforms offering their services to under-16s, plus restrictions on specific functions. The government plans to use the same model as Australia, capturing user-to-user platforms whose purpose is to enable social interaction and that allow users to post material, together with their algorithms. Named platforms include Snapchat, TikTok, YouTube, Instagram, Facebook and X. Messaging services such as WhatsApp and Signal are not intended to be included.

Beyond the ban, the government says it will restrict functions such as livestreaming and stranger communication with children — applying to a wider range of services including gaming sites, and on by default for 16- and 17-year-olds to avoid a “cliff edge.” AI “romantic companion” chatbots must enforce a minimum age of 18. Overnight curfews and breaks in infinite scrolling for under-18s are under consideration, and restrictions on gaming interactions, with detail promised in July. Legislation is expected before Christmas, with first protections in force around Spring 2027.

Sources  
DSIT / Liz Kendall’s statement, 15 June 2026

Does Defend Digital Me support the ban?

No. We support protecting children online, but a blanket ban is the wrong tool. A wide range of children’s rights bodies, safeguarding experts, educators and online safety organisations have reached the same conclusion — among them UNICEF, Child Rights International Network, 5Rights, the Molly Rose Foundation, National Youth Agency, Internet Matters, Girlguiding, teaching unions, Children’s Commissioners from Scotland, Northern Ireland, and Wales, and UN High Commissioner for Human Rights.

The route forward is to address the underlying causes of online harm, build better online spaces, and hold platforms to account for their business models and design — not to exclude millions of children and subject the whole population to age verification.

Sources  
UN OHCHR, Getting children’s safety online right
 · 
UNICEF policy note
 · 
Council of Europe Commissioner for Human Rights

Isn’t the evidence clear that a ban protects children?

The evidence is not one-sided, and it does not all point in one direction. Early findings from Australia, the first country to implement such a ban, show it is not working as intended.

The Molly Rose Foundation’s polling of Australians aged 12–15 found 61% still had accounts on banned platforms, 70% said it was easy to circumvent the ban, 51% said it made no difference to how safe they felt, and 14% reported feeling less safe. The Foundation called UK adoption a “high stakes gamble.” Claiming the evidence points in one direction only works if you decline to look at the rest.

Sources  
eSafety SMMA compliance update, March 2026
 · 
Molly Rose Foundation, Is it working?

Don’t 9 in 10 parents support the ban?

Not as the headline implies. The figure does not mean 90% of UK parents. It comes from the 9,499 parents and carers who chose to answer one consultation question; of those, 89% backed a minimum age. The government’s own notes describe the consultation as a “self-selecting sample” — open to anyone and reflecting people motivated to respond, not a representative national picture.

A self-selecting consultation is a measure of who turned up to answer, not of national opinion — and it tells you nothing about whether the policy works.

Sources  
BBC Verify, “Do nine in 10 parents support a ban?”, 15 June 2026

Which organisations actually back a ban?

Very few in the children’s-rights field. Opposing a blanket ban are UNICEF, the Child Rights International Network, 5Rights, the Molly Rose Foundation, the NSPCC, Internet Matters, Girlguiding, the Back Youth Alliance (ten of the UK’s largest youth organisations), the National Youth Agency, the NAHT and ASCL teaching unions, the Foundation for Information Policy Research, Open Rights Group, Big Brother Watch, and the UN human rights office — alongside three of the four UK Children’s Commissioners. At the Council of Europe’s May 2026 children’s-rights discussion, national governments were the outliers calling for bans; the rights-based organisations opposed it. The weight of expert evidence does not support a ban.

Sources  
Defend Digital Me, “Evidence does not all point in one direction”
 · 
UN OHCHR

Isn’t this a good way to force a pause and get the companies to the table?

No. Politicians should not experiment with children’s lives or punish them for the harms of the providers. Children’s access to information, community and support is not a testing ground for a policy whose own pilot is already failing. The Australian evidence shows the ban does not work: most under-16s keep their accounts or move to unbanned platforms, circumvention is easy, and many report no improvement in safety.

If the mechanism does not deliver the protection it promises, “getting companies to the table” is not a benefit of the ban — it is a justification reached for after the fact. There are direct routes to bringing platforms to account that do not require excluding millions of children and verifying the age of an entire population to find out whether it works. The question is not whether a ban is a clever lever. It is why impose it at all, when the evidence so far says it fails children while costing them their rights.

Won’t children just move to other platforms?

Yes — and Australia demonstrates exactly this. Its ban covers only ten named platforms, which leaves regulators what experts have openly called a game of “whack-a-mole” as ousted under-16s flock to alternatives. On the very day the ban took effect, apps such as Lemon8 (owned by TikTok’s parent ByteDance) and Yope rose up the app-store charts; both had already been put on notice by eSafety.

Within a month, children had also migrated to unbanned sites such as Roblox, Discord and Steam, and a how-to-skirt-the-ban culture emerged. By the end of April 2026, following reports of grooming and extremist gameplay on Roblox, Minecraft, Fortnite and Steam, eSafety had to issue legally enforceable transparency notices to those platforms. Jim Gamble, founding chief executive of the Child Exploitation and Online Protection Centre, says he does not know a single online safety expert who thinks a blanket ban will work, and warns it drives children to “darker corners” of the internet. Even Snapchat warns a ban could push young people onto “less safe platforms.” Deplatforming does not remove harm — it relocates it.

Sources  
ABC News on platform migration (“whack-a-mole”)
 · 
Jim Gamble,ineqe;

Doesn’t the Online Safety Act already deal with harmful content?

It does — and that is precisely the point. The Online Safety Act 2023 already places content restrictions and duties on platforms in law. If the government’s case for a ban is that children are still exposed to harmful content despite the OSA, that is an admission that the OSA has not delivered and that government has not succeeded in regulating large platforms.

The honest conclusion is to make the existing accountability regime work — through Ofcom enforcement, design regulation and the ICO — rather than respond to a regulatory failure by restricting the users instead of the platforms. Reaching for a ban does not fix the enforcement gap; it papers over it, while leaving the harmful business model intact.

Sources  
Ofcom, protecting children under the Online Safety Act

Doesn’t every MP support a ban?

No, they do not. In fact. Labour MP Josh Dean, one of the UK’s youngest MPs, speaking to Politics Home, has said the government’s ban on social media for under-16s is going to “create more problems” for young people rather than making them safer.

Will the government include VPNs in a ban?

No, but messages have been mixed. Josh MacAllister, the Minister for Children and Families, told the BBC Wold At One (15 June, 2026) that next month, the government may set out measures on VPNs, including considering age-gates on VPN use. The House of Lords did indeed debate and originally pass amendments to prohibit children’s access to Virtual Private Networks (VPNs) (specifically Amendment 92) however, after criticism and outrage from civil society, subsequent parliamentary scrutiny led the Government to reevaluate these measures, and remove them from commitments.

Is the underlying evidence on harm as strong as claimed?

It is far more contested than the headlines suggest. One of the most rigorous studies in the field — Orben & Przybylski’s Screens, Teens, and Psychological Well-Being (2019), drawing on time-use-diary data including the Goldilocks dataset of 17,247 adolescents — found little evidence of substantial negative associations between digital-screen engagement and adolescent well-being.

The point is not that social media is harmless — it is that the certainty voiced by many commentators is not matched by the research, which remains largely correlational and disputed. Building a sweeping, hard-to-reverse intervention on a contested evidence base is not a proportionate basis for law.

Sources  
Orben, A. & Przybylski, A. K. (2019), Psychological Science

A note on accuracy: the 91% figure

The likely confusion is with a separate IWF figure that 91% of reported URLs contained some self-generated content; a single such image on a page of a thousand still makes the whole URL count. That is the share of flagged pages, not the share of all imagery. The Minister linked this figure directly to the government’s policy focus and the case for changing the law. Any amount is appalling, but overstating the scale misdirects policy and places an incorrect figure on the permanent record. We have formally requested a correction through the Ministerial Corrections section of Hansard, substituting the accurate figure of 27%.

Sources  
IWF Annual Data & Insights Report 2025
 · 
Hansard, Digital Safety (Children), 8 June 2026

Age verification & biometric data

What are the risks of age verification?

Enforcement requires age assurance, and this is the part the government has not adequately risk-assessed. The crucial point is that age verification applies to everyone, not just children: to prove a user is over 16, every adult must also prove they are not a child. A “children’s” ban becomes population-wide identity infrastructure.

The data-protection record bears this out. Yoti, one of the most-used providers in the Australian ban, was fined €950,000 by the Spanish data protection authority (the AEPD) over its biometric handling (a finding Yoti disputes and is appealing). Worldcoin’s iris-scanning orbs — already live and unmanned in UK shopping centres — have been found in breach of the GDPR by both the Spanish and Bavarian authorities. Risk of breaches is not hypothetical: Discord exposed 70,000 government IDs; journalists linked users of one platform to their driving licences within ten minutes; one TikTok verification partner left photos and licences exposed for nearly a year; and researchers found one provider running 269 separate verification checks, including adverse-media screening, far beyond what users were told.

Sources  
438-signatory joint statement
 · 
AEPD ruling on Yoti
 · 
Discord incident
 · 
AU10TIX / TikTok (404 Media)

“Outrage about face scans is fake — people use Apple Pay all day.”

This is wrong on two counts. First, Apple Pay is a completely different technology from the age verification used by social media platforms. Face ID stays on your own phone: the scan is converted to a mathematical representation held in a secure chip on the device, never sent to Apple, never sent to the shop, never sent to your bank. Nothing about your face leaves your hand. The shop and bank only ever see “this phone authorised a payment.” It is, by design, the opposite of data-sharing.

Age verification usually involves more parties. To prove a user is over 16, the platform — or a third-party verification company acting for it — receives a face scan, a liveness video, or a copy of an identity document, sent off the device into a fragmented chain of commercial processors across jurisdictions, where each transfer is a fresh opportunity for breach, reuse and profiling, and where in some cases the captured data trains commercial AI models. “People use face scans all day” collapses two opposite architectures: one that keeps your biometric on your device and one that ships your face and your papers out to companies you never chose and cannot audit.

There is also a distinct, deeper risk the comparison hides: it normalises children performing for the camera. A liveness check trains a child to turn their face left and right on demand, to prove their body is real, simply to seek permission to talk to friends, play a game or join a local forum. Schools are being warned to take children’s photographs offline because of AI-generated abuse imagery — yet children are simultaneously told to hand over a fresh facial scan for trivial access. Where the check is mandatory and the lowest-friction route is the face, the “consent” a child gives is incentivised and power-imbalanced, not freely given. That is an abuse of consent — and the opposite of what we should be teaching young people about what consent means online.

Sources  
TeaOnHer driving-licence exposure (TechCrunch)
 · 
Open Rights Group on age-verification incentives
 · 
Joint statement of 438 scientists

A child can get around an under-18 alcohol ban too — why not do this anyway?

This was the Prime Minister’s own argument around the announcement: that we don’t abandon an alcohol age limit just because some children evade it. But the analogy breaks on the point that matters. An alcohol age limit does not oblige every adult who looks under 25 to hand over a state-accepted ID that must be captured and retained by a third-party verification company or the platform — as a one-off, or each time. A shopkeeper glances at a card and hands it straight back; nothing is stored, transferred across jurisdictions, or fed into a model. For those who look over 25, they don’t even come into scope. Everyone is in scope in this social media ‘minimum age requirements’ ban.

Online age verification, to prove a user is over 16, means asking everyone to submit an ID, and often a face scan or image is fed into a commercial chain that records the transaction, retains the data for set periods, and in some cases reuses it. The offline comparison the government reaches for is precisely the one that exposes what is different and disproportionate about the online version: a momentary human glance versus a permanent, transferable, breachable digital record of your face or your papers.

Sources  
PM’s remarks on the announcement, BBC News, 14–15 June 2026

Why does biometric data matter so much more than a password?

Because it cannot be reset. A breached password can be replaced; a scanned face or iris cannot. As the Bavarian regulator put it in the Worldcoin case, once a company holds your iris pattern you can never be anonymous to it again. For every child who has already handed the data over, the harm is permanent.

This is sharpened by where the technology is heading. The same platforms now legally required to hold children’s biometric data are wiring AI agents into their account systems. In the 2026 Instagram incident, attackers manipulated Meta’s AI-powered account-recovery assistant into resetting passwords simply by sending it a natural-language instruction — no credential theft, no server breach. Age verification concentrates exactly the data that makes such a failure irreversible: a leaked face scan or identity document cannot be reissued. General Comment No. 25 calls on states to ensure any biometric processing of children is necessary, proportionate and accompanied by strong safeguards — a test mandatory age verification does not meet.

Sources  
BayLDA Worldcoin decision
 · 
UN Committee on the Rights of the Child, General Comment No. 25

Haven’t curfews and “shutdown laws” been tried before?

They have, and they failed. In 2011 South Korea introduced its “Shutdown Law” (the so-called Cinderella Law), blocking under-16s from games websites after midnight. Ten years later, in 2021, the country abandoned it — it had proved unworkable and unpopular, and research found it increased, rather than reduced, young people’s online hours, with negligible effect on sleep or academic performance.

The UK’s stated interest in overnight curfews and scrolling breaks for under-18s revisits an idea another country has already road-tested to destruction over a decade. Repeating a discredited policy is not learning from the evidence; it is ignoring it.

Sources  
CNN (2011)
 · 
NME (2021), South Korea abandons the curfew
 · 
Lee, Kim & Hong (2017)

Children’s rights

UNCRC Article 13 · GC25

What does a ban do to children’s right to access news and information?

It cuts it off for those it actually excludes. Article 13 of the UN Convention on the Rights of the Child protects the child’s right to seek, receive and impart information; General Comment No. 25 makes clear this applies fully in the digital environment, and that children rely on digital access to exercise it. Social media is now a primary route by which young people encounter news.

This is a direct rights cost that has been largely absent from the UK debate — and it runs contrary to the claim, made in the 8 June debate, that British children’s access to news would be unaffected. The Australian evidence already shows otherwise.

Sources  
Notley et al. (2026), young people’s news engagement
 · 
General Comment No. 25

UNCRC Article 31 · GC25

What about children’s right to play?

Article 31 protects the child’s right to rest, leisure, play and participation in cultural life. General Comment No. 25 explicitly extends this to the digital environment, recognising that play, creativity and social connection increasingly happen online, and that opportunities to realise these rights digitally should be expanded, not removed.

Sweeping gaming and “stranger communication” functions into the restrictions — as the announcement signals for a wider range of services including gaming sites — directly engages this right. Australia’s experience shows children displaced from named platforms simply moved onto gaming and chat services such as Roblox, Discord and Steam, so the restriction neither protects them nor preserves the play and connection those spaces provide. A rights-based approach treats safe digital play as something to design for, not something to gate behind an age check.

Sources  
General Comment No. 25 (play and leisure)
 · 
Council of Europe, Educating for a video game culture

UNCRC Article 16 · GC25

How does the ban affect children’s right to privacy?

It inverts it. Article 16 protects the child against arbitrary or unlawful interference with privacy. General Comment No. 25 calls on states to ensure that children’s privacy is respected in the digital environment, that any interference is lawful, necessary and proportionate, and that biometric processing of children carries strong safeguards.

Mandatory age verification does the opposite: it requires children to surrender high-value, sensitive biometric data — face scans, liveness checks, sometimes identity documents — simply to get online, passed through commercial processors and, in Meta’s case, into the training data for generative AI. Teaching children to trade their bodily data for everyday participation, under coerced “consent,” is incompatible with Article 16 and with what General Comment No. 25 asks of states. It is also the opposite of the message children should learn about what consent means online.

Sources  
General Comment No. 25 (privacy)
 · 
AEPD ruling on Yoti’s biometric handling

UNCRC Article 12 · GC25

Have children been heard on a policy made in their name?

Barely. Article 12 gives children the right to express their views freely on all matters affecting them, and to have those views given due weight. This policy plainly affects children — it restricts where they may go, speak and seek support online — so Article 12 is directly engaged. General Comment No. 25 treats hearing children as a precondition of any valid assessment of their best interests in the digital environment.

Yet a junior minister signalled, before the consultation had even closed, that “under any outcome we will impose some form of age or functionality restriction on children under 16.” A decision presented as settled before children’s views were gathered, let alone weighed, skips a step the Convention treats as essential. Where young people have been asked, only 15% of 10–16-year-olds thought a ban would make them feel safer. Shutting young people out of a decision made in their name risks their trust that politicians represent them at all — at the very moment the voting age is moving to 16.

Sources  
General Comment No. 25 (right to be heard)
 · 
Girlguiding research, March 2026

The Children’s Commissioners support this — don’t they?

Not the ones closest to children’s own evidence. Three of the UK’s four Children’s Commissioners oppose a blanket ban. Scotland’s Commissioner carried out a children’s rights impact assessment and concluded against one. Northern Ireland’s Commissioner warns that over-simplistic bans “can, in fact, make things worse by pushing them into more dangerous, unregulated online spaces.” Wales’s Commissioner, after engaging children across Wales, says removing children from digital spaces “is not the answer,” and that access should be “conditional and earned” by platforms proving they are safe. Only England’s Commissioner has called for a ban — and alone, up to age 18.

Sources  
Children’s Commissioner for Wales, Growing up in the online world
 · 
Children & Young People’s Commissioner Scotland

What now

What would actually help?

Tackle the root causes rather than the users. The more effective route — as Australia’s own Human Rights Commissioner and National Children’s Commissioner argued before that country’s ban — is a legal duty of care that compels platforms to make their environments safe: moderating harmful content and preventing predatory behaviour. The 438 signatory scientists made the same point: most of the harms age checks are meant to address are caused by the algorithmic practices of social networks, so regulating those practices is more directly effective than circumventable access controls.

The aims of the Online Safety Act shifted over time, from systemic change toward content moderation — a move civil society has repeatedly warned, since 2015, is a mistake. We should go back toward improving the environment for children, not removing them from it. If the government cannot enforce its own law, it needs to address platform power — using competition law to break up platforms, for example: it is unclear why a single company should be able to own multiple social networks, when the size and reach of conventional media are regulated.

In practice, at platform level for a safer environment: regulate platform design and business models; ban addictive design features aimed at children; enforce existing data-protection law; resource Ofcom and the ICO to make the Online Safety Act actually bite; support parents with local controls; and invest in digital literacy. Restoring the mental-health provision and youth services that have been cut would do more for children’s wellbeing than blocking access — when nearly 40,000 children wait more than two years for mental-health support, and youth-service funding has fallen by 76% in real terms over 14 years, a ban addresses neither.

Sources  
Australian Human Rights Commission op-ed
 · 
UNICEF policy note

What happens next?

The government intends to legislate before Christmas, using secondary legislation under the Children’s Wellbeing and Schools Act, with first regulations expected in force around Spring 2027. Ofcom has been asked to conduct a rapid study on “highly effective age assurance” for verifying who is over 16, and to review its enforcement capabilities. Separately, the Online Safety Act 2023 requires Ofcom to publish a report by 17 July 2026 assessing how services have used age assurance and how effective it has been — it would be premature to mandate policy before that evidence is public.

Sources  
Defend Digital Me research

Isn’t each new measure at least progress?

Not according to the pattern. Open Rights Group describes a “policy ratchet”: for over a decade each child-safety measure — default ISP filters, then the Online Safety Act, now a ban — has delivered far less than promised, and each failure has prompted a more extreme next step rather than a rethink. The measures fall on users, in age checks and privacy and free-expression costs, while leaving the engine of harm — algorithmic amplification and the attention-based business model — untouched. ORG argues the effective interventions act on the platforms instead: user switching rights, letting users control their own feeds, and competition remedies. A ban is the next turn of the ratchet, not a break from it.

Sources  
Open Rights Group, “The social media policy ratchet”, 15 June 2026

Defend Digital Me

Children’s data rights in education and online. This briefing may be shared and reused under CC BY-SA 4.0, with attribution.

Contact: jen@defenddigitalme[dot]org  ·  defenddigitalme.org