UK's Under-16 Ban Copies Australia's 70% Failure. Proton Shows What That Means.

The 70% Stat: Australia's Under-16 Ban Didn't Catch the Target Cohort

Australia's under-16 social media ban took effect in December 2025. The architecture was simple. Platforms were required to use "reasonable steps" to verify ages, including government ID, credit card checks, and third-party age-estimation services. The Australian government sold the ban as a child-safety intervention, and the political support for the policy was broad enough that the UK government is now importing the template.^[1]

Six months of data are now in. The results are not encouraging.

Australia's eSafety regulator found that 70% of under-16s in the country continue to access banned platforms.^[1] Teens are bypassing the restrictions by providing false credentials at account sign-up or lying about their age. The workarounds are the same ones that existed before the ban. The age-verification infrastructure caught something. It did not catch the target cohort.

The structural read is the one Proton's analysis surfaces: an age-verification regime that catches 30% of its target cohort forces the rest to either lie about their age or adopt a work-around. The lie-at-sign-up path is the dominant one. The 70% figure is the shareable civil-liberties angle. The Starmer package is the same template, applied to a country six times Australia's population, with a more aggressive timeline.

The Ofcom VPN Data: Children Aren't Driving the Surge. Privacy Is.

Australia's eSafety guidelines also point to VPNs as a circumvention tool and ask platforms to detect and block them. The framing assumes children are the primary VPN users, since the regulatory point of age verification is to keep children off the platforms.^[1]

The data does not support the framing.

When the UK's Online Safety Act brought in age verification requirements in 2025, VPN usage more than doubled. Ofcom's Online Nation report found the surge had fallen back significantly by November 2025. Research from Childnet, the UK children's internet-safety charity, found the surge was not attributable to children at all. The most common reason children gave for using a VPN was to stay safe online and protect their privacy.^[1]

That is the privacy-as-self-defense angle. The children who are using VPNs are not doing it to bypass a social media ban. They are doing it to protect themselves from the data-collection infrastructure that the same regulatory regime is building. The Proton piece surfaces the point: the assumption that "more age verification" reduces the privacy harm to children assumes the children are the threat model. The Ofcom data says the children are the threat-model target.

The US Government Submission: VPN Restrictions "Are Typically Associated With States That Subject Their People to Significant Censorship"

The third data point Proton surfaces is the one with the longest reach. The US government submitted a formal objection to the UK consultation on the Starmer age-verification package. The objection, in the words of the submission, is plain.^[1]

VPNs are a useful, lawful privacy tool. Individuals globally rely on VPNs as an essential tool to protect their privacy online and access the open internet. Policies banning or treating such internet freedom and privacy tools as inherently suspect are typically associated with states that subject their people to significant censorship and human rights violations.^[1]

The US objection is the international / diplomatic angle. The same regulatory logic that bans under-16 social media use, when extended to block the workarounds, treats VPNs as a circumvention tool. The US government is telling the UK that treating VPNs as a circumvention tool is structurally identical to the censorship logic the US has spent the last decade criticizing. The diplomatic framing matters because the Starmer package is expected to be unveiled in the King's Speech in the autumn. The US submission is on the public record before the unveiling.

The Proton piece notes that the UK has been here before. The government spent years trying to introduce a system requiring users to prove their age to access pornography online. The effort collapsed in 2019 after repeated technical failures and the discovery that at least one verification system could be bypassed in minutes.^[1] When the government tried again under the Online Safety Act, the results were no more reassuring: Discord's rollout of age verification in 2025 ended with a third-party provider compromised and 70,000 government-ID photos exposed.^[1] The age-verification infrastructure has a documented track record of failure on the privacy side.

The Starmer Package: What's Being Built on Top of the 70% Failure

Prime Minister Keir Starmer announced on Monday, June 15, 2026 that the UK will ban children under 16 from social media platforms including Instagram, TikTok, YouTube, Facebook, Snapchat, and X. Legislation is expected before Christmas. The ban is scheduled to come into force in spring 2027. Messaging services like WhatsApp and Signal are excluded.^[1] The UK is going further than Australia in some respects: the ban will extend to livestreaming and stranger-to-child contact on gaming platforms, and the government is considering overnight curfews on social media use for under-18s. AI chatbots designed to simulate romantic relationships will be restricted to adults only.^[1]

Political support is broad. More than 90% of parents who responded to the government's public consultation backed a minimum age of 16. Starmer framed the decision in unambiguous terms: "Tech giants had their chance and failed, but we're stepping in to protect children."^[1] The 90% figure is real, and the child-safety case is the political case. The privacy cost is the cost the public has not been asked to weigh in on.

The Proton piece is direct about the cost. Every age-verification system, however it works, requires platforms to collect more personal data than they do today. The UK government has asked Ofcom to conduct a rapid study on what constitutes "highly effective age assurance" for verifying whether someone is over 16.^[1] That study will likely revisit the same territory already explored in Australia: biometric facial-age estimation, live selfie verification, AI-based behavioral inference, and government-issued identity document uploads.^[1]

The Proton piece is direct about the result, too. As Proton noted when Australia's ban took effect, this approach turns mainstream social platforms into identity-verified services rather than places where people can participate without handing over sensitive information.^[1] The framing is the privacy hook for the Starmer beat. The announcement is the child-safety case. The implementation is the privacy case. The two are bundled, and the public consultation asked about the first one.

EFF: The Provisions the Starmer Announcement Didn't Surface

The Electronic Frontier Foundation published a 4,200-word analysis on June 19, 2026, four days after the Starmer announcement, that goes further than the Proton analysis on four structural points the announcement-driven coverage did not have.^[3]

First, the legislative vehicle. The Starmer package is not a standalone bill. It is a clause in the Children's Wellbeing and Schools Bill. In March 2026, "politicians introduced plans for a social media ban into the Children's Wellbeing and Schools Bill to 'prevent children under the age of 16 from becoming or being users' of 'all regulated user-to-user services,' to be implemented by 'highly-effective age assurance measures'."^[3] The Proton piece surfaced the architecture. The EFF piece surfaces the legislative vehicle: the same political-economy logic the Australian Online Safety Act used to import age verification, bundled with school-wellbeing funding, which makes it harder to vote against.

Second, the US-academic influence is on the record. Jonathan Haidt, the social psychologist whose 2024 book "The Anxious Generation" is the canonical US case for an under-16 social media ban, "met with the UK Health Secretary in February to push for the ban," per the EFF.^[3] The EFF piece names the scientific-doubt context around Haidt's research base directly: "American social psychologist and The Anxious Generation author Jonathan Haidt, who has called for age-related social media bans around the world, despite significant scientific doubt about his research, met with the UK Health Secretary in February to push for the ban." The US-academic-as-policymaker-import path is structural: an academic with a contested research base crosses the Atlantic, briefs a minister, and a regulatory template is built on top of the academic's preferred intervention. The Proton piece does not name Haidt. The EFF piece does.

Third, the age-18 amendment path. "When this proposal came before the House of Commons, MPs defeated and proposed their own amendment: enabling the Secretary of State to introduce provisions 'requiring providers of specified internet services' to prevent access by children, under age 18 rather than 16, to specified internet services or to specified features; and to restrict access by children to specified internet services which ministers provide."^[3] The amendment makes the age ceiling a ministerial discretion, not a statutory floor. The Starmer announcement described a 16+ ban. The EFF piece surfaces that the legislative text permits the Secretary of State to set the ceiling higher, with secondary legislation, after the bill passes. The announcement and the statute are not the same number.

Fourth, the ISP-level provisions. "The provision also requires internet service providers to limit the time kids spend online, and has rules about who can contact them online. These extreme rules will take decisions about using technology away from families and put them in the hands of government regulators," the EFF piece says.^[3] The Proton piece focused on age verification at the platform layer. The EFF piece surfaces that the regulatory template extends past the platform to the ISP layer. ISP-level enforcement is harder to circumvent than platform-level enforcement: a VPN gets past a platform's age gate, but it does not get past an ISP-level time limit. The regulatory layer the EFF piece names is the layer Proton did not.

The EFF framing of the whole package is that the ban will cause more harm than it prevents, and the EFF draws the line back to the Australian template. "The history of this proposal shows that the UK government has repeatedly returned to the same flawed idea: restricting access to online services by requiring age checks for everyone," the EFF piece says. "But the fundamental problems have not changed. There is still no widely available way to verify age online without compromising privacy, but even if there were, broad restrictions on social media will inevitably limit access to lawful speech, and valuable online communities, and arts and culture."^[3] The age-gating history the EFF piece traces is also load-bearing: the Digital Economy Bill a decade ago, the Digital Economy Act of 2017, the 2020 Online Harms Whitepaper, the 2023 Online Safety Act, the July 2025 age-assurance implementation on harmful-content sites, and the March 2026 Children's Wellbeing and Schools Bill insertion. The same regulatory template, in six acts of parliament over a decade, against the same age-verification problem that the EFF piece says has no reliable, privacy-preserving solution.^[3] The Proton piece's Ofcom "highly-effective age assurance" study sits on top of the same decade-long regulatory continuum the EFF piece documents.

The HN Discourse: 210,576 Petition Signatures, the Canadian Federal Parallel, and the Zero-Knowledge Counter-Proposal

The EFF piece hit Hacker News on June 19, 2026 at 21:58 UTC, submitted by HN user Brajeshwar. The thread crossed 68 points with 74 comments by 08:38 UTC June 20, a 1.09 comment-per-point ratio that puts the thread in the high-discourse-density band for HN surveillance and privacy threads.^[4] The substantive voices concentrated on the implementation-mechanics side rather than the ban-yes-ban-no side. Four HN comments are load-bearing for the structural critique.

First, the structural implementation-mechanics argument that earned the most upvote weight came from HN user big85. "About 2 million adults in the UK don't have government-issued photo ID. Certainly many 16-17 year olds will have trouble verifying their age. They're blocking huge sectors of the UK population from being able to use the internet normally."^[4] The 2 million figure is the load-bearing implementation-mechanics argument against the Starmer package. The Starmer announcement assumes age verification is technically possible for everyone under 18. The UK government does not have a national photo-ID database. The 2 million figure is the share of the adult population the Starmer package will lock out of the regulated platforms, on top of the under-18 ban itself.

Second, the Canadian federal-government parallel surfaced from HN user btbuildem. "Canadian federal govt just pushed thru a slew of similar legislation, absolutely unprecedented assault on privacy, tools for tracking everyone all the time, minimally constrained, giving broad leeway to a three-person unelected body to implement the actual details."^[4] The Canadian parallel is structurally important: the Starmer package is not a UK-only regulatory template. Five Eyes jurisdictions are landing on the same age-verification-and-time-limit regulatory architecture in the same regulatory window. The regulatory template is converging across jurisdictions before the public has data on whether the Australian template worked.

Third, the technical counter-proposal surfaced from HN user cbdumas. "If governments want to set up online age gates, they should be responsible for providing an electronic ID system and enable privacy preserving age verification (zero knowledge proofs)."^[4] Zero-knowledge proofs are the cryptographic technique that lets a user prove a property (over 16) without disclosing any other information (real age, real identity). The ZKP approach is the technical alternative the EFF piece does not name but the structural critique points at. The Proton piece's Ofcom "highly-effective age assurance" study is the regulatory hook for the ZKP alternative. If Ofcom's recommendation names ZKPs as compliant, the Starmer package can be re-engineered without the platform-side data-collection architecture the Proton piece critiques.

Fourth, the UK Parliament Petition 757233, surfaced in the HN thread by user throw5, is the formal parliamentary channel for the regulatory opposition. The petition, "Do not ban social media for under 16s," was created on June 15, 2026 by Leo Rhodes, crossed 100,000 signatures the same day, and is now at 210,576 signatures as of June 20, 2026. The Petitions Committee confirmed on June 15, 2026 that it will be considered for debate in Parliament.^[5][4] The petition is the institutional hook for the regulatory opposition. The Proton piece did not name it. The EFF piece does not name it. The HN thread is the only public record that ties the petition to the EFF structural critique, and the petition is the only public record that links the Starmer package to a formal UK Parliamentary debate with the regulatory opposition on the schedule.

Watch for the Petitions Committee debate date. The Petitions Committee has confirmed it will be debated. The date is not yet on the public record. When the debate is scheduled, the EFF structural critique lands on the UK Parliament floor, on the public record, alongside the Starmer announcement.

What the Existing Starmer Australia-Plus Coverage Doesn't Say

The State of Surveillance coverage of the Starmer "Australia plus" under-16 ban landed on June 15, 2026. The piece is anchored on the Times, Guardian, and Independent reporting on the announcement, plus the Signal warning that the UK CSAR / Online Safety encrypted-device-scanning mandate "endangers us all."^[2]

That piece was right about the regulatory continuum: the under-16 ban, the AI chatbot restriction, and the CSAR mandate are three interventions sold as one child-safety agenda. The Proton piece published the same day surfaces the data points the announcement-driven coverage did not have. The 70% stat. The Ofcom VPN data. The US government submission. Three privacy hooks, all in one analysis, all on the public record the day of the Starmer announcement.

The privacy story is the implementation gap, not the announcement. Australia's ban has been live for six months. 70% of the target cohort still gets in. The UK is about to build the same infrastructure, on a larger scale, with a more aggressive timeline, in a country with a documented prior failure of the same technology, in a regulatory environment where the US government has formally objected to treating VPNs as a circumvention tool. The Proton analysis is the first major public write-up that puts the data points in the same paragraph. The privacy angle is the angle the regulatory push has been trying to avoid being defined by.

The Critics: "It Protects Platforms, Not Children"

The Proton piece surfaces the criticism the Starmer package has not yet engaged with. The Molly Rose Foundation, established in memory of 14-year-old Molly Russell, who died by suicide after viewing self-harm content online, argues the ban is too blunt an instrument.^[1]

"What we're really concerned about is that the government rushes into solutions that the evidence just doesn't support, rather than addressing the causes of harm," said Rowan Ferguson, the foundation's policy manager. Kate Edwards, its head of education, put it more directly: "It does nothing to address the actual problem. The harmful algorithms, the harmful content. The ban doesn't touch the business models."^[1]

The Proton piece lands on the same line the critics land on. The business models that make the platforms harmful, the algorithmic amplification of extreme content, the infinite scroll, the engagement-at-all-costs design, affect adults just as much as children. A ban on under-16 access does not touch those models. Even if it removes some young people from the platforms, it does not make the platforms any less toxic. The more lasting solution is to make the internet less harmful for everyone.^[1]

The Proton piece closes with the framing. A ban on under-16s protects platforms, not children.^[1] The framing is the shareable civil-liberties angle the announcement-driven coverage has not yet integrated. The data points are the proof. The proof is what the Starmer package has to answer to, not the announcement.

What It Means for You Today

Three groups are affected, and the consequences diverge.

If you are a UK-based user of social media or AI chatbots. Nothing has changed yet. The package has not been published. The King's Speech is expected in the autumn. The implementation timeline, on the Proton piece's account, is spring 2027 for the under-16 social media ban, with the AI chatbot restriction on a parallel track.^[1] Expect the same compliance patterns the Australian template produced: most platforms will build the age-verification infrastructure, some will block UK under-16 traffic entirely, and a small number will exit the UK market. The compliance cost falls on the platforms. The data falls on the user.

If you are a UK-based user of a VPN. The Proton piece surfaces a real risk. The Starmer package treats VPN detection as a circumvention-control requirement. Australia's eSafety guidelines already ask platforms to detect and block VPNs.^[1] The UK Ofcom study on "highly effective age assurance" is likely to land in the same regulatory territory. The US government submission is the diplomatic counter-argument: VPNs are a lawful privacy tool, and treating them as a circumvention channel is structurally identical to the censorship logic the US has spent a decade criticizing.^[1] Watch for the Ofcom study's specific recommendation on VPNs. That recommendation is the operational test of whether the Starmer package accepts or rejects the US objection.

If you are a parent of a UK-based under-16, or a UK-based under-16 yourself. The package is sold as child-safety legislation. The child-safety case is real. The package, if implemented as described, will require every social media platform and every AI chatbot provider to verify the age of every UK user, on sign-up and on an ongoing basis. The verification, by design, requires the user to disclose information they would not otherwise disclose. The trade-off is not child safety versus privacy. The trade-off is one form of child safety (age-gated access to platforms) versus another (the privacy of the child's data, the security of the child's identity documents, and the integrity of the child's interaction with every service that touches the age-verification infrastructure). The package does not ask the public to make that trade-off explicitly. The trade-off is buried in the technical implementation, and the implementation is the part the public consultation did not cover.

The Bottom Line

Australia's under-16 social media ban has been live for six months. 70% of Australian under-16s still access banned platforms.^[1] The UK's Online Safety Act age-verification push correlated with a more-than-doubling of UK VPN usage, but Ofcom's Online Nation report and Childnet research found the surge was not driven by children, the most common reason kids gave for VPN use was to stay safe online and protect their privacy.^[1] The US government told the UK consultation that VPN restrictions are "typically associated with states that subject their people to significant censorship and human rights violations."^[1]

Keir Starmer is preparing to import the Australian template. The Proton analysis, published the same day as the Starmer announcement, is the first major public write-up that puts all three data points in the same paragraph.^[1] The story is the implementation gap, not the announcement. The Starmer package has to answer the 70% stat, the Ofcom VPN data, the US submission, and the Discord age-verification compromise from 2025, all on the same public record. The privacy hook is the angle the regulatory push has been trying not to be defined by. The data is in. The angle is set.

Watch for three things over the next 90 days. First, the publication of the King's Speech in the autumn, and the specific text of the under-16 social media ban and the AI chatbot restriction. Second, the Ofcom "highly effective age assurance" study, and the specific recommendation on VPNs. Third, the UK government's response to the US government submission, on the public record, before the package is finalized. The three answers, taken together, will determine whether the Starmer Australia-plus package is the child-safety policy the consultation claimed it was, or the privacy-degradation vector the Proton data points describe.

Sources

1. Proton: "UK's social media ban for children: the privacy problems Australia already exposed" (Edward Komenda, June 15, 2026, primary source for the eSafety 70% stat, the Ofcom Online Nation report citation, the Childnet research finding on the most common reason children give for VPN use, the US government submission to the UK consultation with direct quote, the Starmer June 15 announcement details, the prior UK age-verification failure history including the 2019 collapse and the Discord 2025 compromise, the Molly Rose Foundation critique, and the closing "protects platforms, not children" framing)
2. State of Surveillance: "Starmer Is Importing Australia's Age Verification Regime. Then Going Further." (June 15, 2026, the prior UK Starmer Australia-plus coverage, anchored on the Times / Guardian / Independent reporting and the Signal CSAR warning; the Proton piece is the privacy follow-up that integrates the eSafety 70% stat, the Ofcom VPN data, and the US government submission into a single canonical reference)
3. EFF Deeplinks: "The UK's New Under-16 Social Media Ban Will Cause More Harm Than It Prevents" (Paige Collings and Jillian C. York, June 19, 2026, primary source for the structural critique beyond the Proton piece: the Children's Wellbeing and Schools Bill insertion in March 2026 with the "highly-effective age assurance measures" enforcement framework, the Jonathan Haidt February 2026 UK Health Secretary meeting with the "despite significant scientific doubt about his research" framing, the House of Commons amendment enabling the Secretary of State to raise the cap from 16 to 18 in secondary legislation, the ISP-level time limits on kids' online time and the stranger-contact restrictions with the "extreme rules will take decisions about using technology away from families and put them in the hands of government regulators" framing, the "no reliable, privacy-preserving method of verifying the age of every internet user" framing, the EFF "returned to the same flawed idea" historical framing from the 2017 Digital Economy Act to the 2025 Online Safety Act age-assurance implementation, and the EFF "broad restrictions on social media will inevitably limit access to lawful speech, and valuable online communities, and arts and culture" closing framing)
4. Hacker News thread on EFF Deeplinks article "The UK's New Under-16 Social Media Ban Will Cause More Harm Than It Prevents" (HN id 48603761, submitted by Brajeshwar at 2026-06-19T21:58:22Z, 68 points and 74 comments at 08:38 UTC June 20, 1.09 comment-per-point ratio in the high-discourse-density band for HN surveillance and privacy threads, four load-bearing HN comments: big85 on "About 2 million adults in the UK don't have government-issued photo ID" as the structural implementation-mechanics argument, btbuildem on the Canadian federal-government parallel as the five-eyes-jurisdictional-convergence data point, cbdumas on zero-knowledge-proof age verification as the technical counter-proposal, throw5 linking to UK Parliament Petition 757233 as the formal parliamentary channel for the regulatory opposition, enoeht on decentralized social media workarounds as the political-economy angle, drawfloat on the ISP-level time-limit provisions)
5. UK Parliament Petition 757233: "Do not ban social media for under 16s" (created by Leo Rhodes on June 15, 2026, 210,576 signatures as of June 20, 2026, Petitions Committee confirmed on June 15, 2026 for debate in Parliament after the petition crossed the 100,000-signature threshold the same day, closes August 11, 2026, the institutional hook for the regulatory opposition, linked to the House of Commons Library research briefing "Proposals to ban social media for children" dated March 19, 2026 and the Science, Innovation and Technology Committee evidence session "Social media age restrictions" dated March 11, 2026 and the Department for Science, Innovation and Technology consultation "Growing up in the online world" with deadline May 26, 2026)

Related Coverage

• Proton Mail review. Our assessment of the provider quoted here
• Starmer Is Importing Australia's Age Verification Regime. Then Going Further. The June 15, 2026 piece on the Starmer "Australia plus" announcement, the Guardian "Australia plus" framing, the AI chatbot restriction, and the Signal CSAR "endangers us all" warning. The Proton piece published the same day is the privacy follow-up: the 70% stat, the Ofcom VPN data, and the US government submission are the three data points the announcement-driven coverage did not yet have.
• Age Verification Triggers VPN Crackdown Across Three Continents The March 25, 2026 piece on the prior Australia / UK / US age-verification and VPN crackdown cycle. The Proton June 15 piece confirms the cycle has now produced real-world data on the privacy consequences: a more-than-doubling of UK VPN usage, and the Childnet finding that children are not the primary driver of the surge.
• Australia's Age Verification VPN Surge and the Privacy Dilemma The March 19, 2026 worked-example reference for the Australian template. The age verification rules triggered a sustained increase in VPN downloads, with the Proton piece now confirming the post-November 2025 drop-off. The Starmer "Australia plus" package imports the same template and should expect the same compliance response.
• Age Verification Is the Surveillance Infrastructure the ID System Was Always Going to Build The structural read on age verification as the load-bearing infrastructure for every "child-safety" use case. The Proton piece confirms the structural prediction: the Ofcom "highly effective age assurance" study is now commissioning the same architecture Proton warned about when Australia's ban took effect.
• Yoti, GrapheneOS, and the Phone-Reported-to-Authorities Privacy Line The June 10, 2026 piece on the dominant UK age-verification vendor's data-handling practices. Yoti is the vendor the Australian eSafety framework points platforms toward, and the Yoti infrastructure is the same one the UK Ofcom "highly effective age assurance" study is likely to evaluate.
• Discord Age Verification, 70,000 Government-ID Photos, and the Biometric Privacy Compromise The Discord age-verification compromise Proton cites in the June 15 piece. The 70,000 government-ID-photo exposure is the prior data point for the Starmer package's "highly effective age assurance" study, and the precedent for what the UK Ofcom evaluation is supposed to prevent.
• Utah's VPN Law, SB 73, and the Age-Verification Effective Date The US state-level analogue. Utah's SB 73 took effect May 6, 2026 and treats VPN use as an aggravating factor in age-verification enforcement. The US government submission to the UK consultation is the diplomatic parallel: the same regulatory logic, in the same regulatory window, on the same VPN-circumvention question.
• Wisconsin, Michigan, and the State-Level VPN Ban Wave The US state-level follow-up to the Utah VPN law. The Proton piece's US government submission is the federal-level parallel: the same "VPNs as circumvention tools" framing, at the federal regulatory level, in the same regulatory window as the Starmer package.
• Big Tech Government Data Requests Surged 770% Since 2014. Proton Has the Receipts. The Proton transparency report that establishes Proton's standing as a privacy-by-default vendor. The June 15 Proton piece is the editorial counterpart: the same privacy-default framework, applied to the Starmer package, anchored on the eSafety 70% stat and the US government submission.
• Australia's ACMA SMS Sender ID Registration The June 18, 2026 Australian Communications and Media Authority announcement requiring SMS/MMS sender ID registration, the morning cycle's fastest engagement compound per the Hacker News 86 points and 46 comments, and the same Australian regulatory template applied to the SMS channel with the Ofcom UK parallel closing the cross-border enforcement gap.
• Australia Surveillance Laws 2026: The Full Picture The architecture overview. The Assistance and Access Act 2018 (TAA), the Digital ID Act 2024, the eSafety Commissioner under the Online Safety Act 2021, and the AUSTRAC / ASIO / TIA Act 1979 surveillance authorities underneath the under-16 ban. The Proton 70% stat and the Ofcom VPN data sit on top of the same Australian template that this piece imports.
• Kids Online Safety Act (KOSA): Protecting Children or Censoring the Internet? The US federal-level parallel for the Starmer Australia-plus package. KOSA's "duty of care" framework requires platforms to protect minors, and the EFF and civil-liberties pushback on KOSA is the US domestic counterpart to the EFF's UK critique. The EFF's June 19, 2026 UK piece draws on the same age-verification-and-censorship framing the EFF and CDT have used against KOSA since 2023.
• California Wants Your ID to Scroll Instagram. The Privacy Fallout Is Massive. The California state-level parallel. AB 1709 requires every California user to verify their identity with government ID or biometrics, the same "highly-effective age assurance" regulatory architecture the EFF piece critiques in the UK Children's Wellbeing and Schools Bill. The EFF's alarm on AB 1709 is the same alarm the EFF raises on the Starmer package.