Don’t Be Fooled: What Australia Calls a Social Media Delay Is Actually a Triple Ban

The word “delay” is doing a great deal of work right now. Not only in Australia, but also in Brussels.

On 12 May 2026, European Commission President Ursula von der Leyen addressed the European Summit on Artificial Intelligence and Children in Copenhagen. She called for action. She cited platform addiction, algorithmic harm, and the exposure of children to distressing content. The speech was substantive, and it landed. But it also contained a word choice that deserves attention. “Without pre-empting the panel’s findings,” she said, “I believe we must consider a social media delay.”

A delay. The same word Australia’s eSafety Commissioner has used since December 2025 to describe a law that Australia’s own national dictionary centre chose “social media ban” to describe as its Word of the Year. Commissioner Julie Inman Grant told the UK House of Commons Science, Innovation and Technology Committee that her office refers to it as “a social media delay, because it isn’t an absolute ban.” She has repeated that framing consistently since.

The public understood what the Commissioner did not quite say. And von der Leyen in Europe, following Australia’s blueprint, is now reaching for the same language.

The word “delay” implies a pause. It suggests that what is being withheld is being held in trust, to be returned in full at a later date. It is a framing that forecloses scrutiny by making the policy sound temporary, benign, and reversible. If access is merely delayed, then nothing permanent has been done. Nothing has been lost.

That framing does not survive contact with what the law actually does. Australia’s experiment, now five months old, shows that three things are being restricted. One of them is temporary. Two are not.

A Triple Ban That “Delay” Cannot Describe

A Ban on Anonymity for Everyone

Age gating and/or age verification does not touch only minors. It touches everyone.

To establish that a user is 16 or older, platforms must collect and verify identity information. The methods under discussion and in use include facial age estimation, biometric checks, government identification, credit card linkage, and third-party intermediary tokens. None of these are neutral. Each requires a user to submit data about who they are in order to access a service that, previously, required no such disclosure.

The result is structural: online anonymity, which has historically served not as a shield for bad actors but as a condition for ordinary expression, political dissent, survivor testimony, and minority participation, is eliminated as a precondition of social media access. This applies to every adult, not only those under 16. Age verification, by design, knows no minimum. It checks everyone to exclude some.

Privacy experts, digital rights organisations, and Proton, among others, have noted that this shift is not incidental to the policy. It is the policy. Age gating converts social media platforms from spaces of pseudonymous connection into identity-verified services. The implications extend beyond convenience.Age verification systems have already produced significant data breach events, including AU10TIX in 2024, Discord/Zendesk in 2025, and Persona in 2026. The cost of being known online is borne not only by the young people the law claims to protect, but by the entire population whose anonymity is the collateral. That cost is a function of how robust the verification infrastructure is.

Von der Leyen went further in Copenhagen. She told delegates that “the technology for age verification is available,” and pointed to the EU’s own age-verification app, modelled on the EU Digital COVID Certificate, as proof. She had launched it less than a month earlier, on 14 April 2026, describing it as meeting “the highest privacy standards in the world” and ticking “all the boxes.” On 16 April, UK security consultant Paul Moore bypassed it in under two minutes by editing a plain-text configuration file on an Android device. He addressed von der Leyen directly, warning that “this product will be the catalyst for an enormous breach at some point. It’s just a matter of time.” Researchers confirmed additional flaws: rate-limiting controls stored in an editable file, biometric authentication disableable with a single boolean switch, and a separate architectural flaw identified in March 2026 whereby the system cannot verify that passport validation actually occurred on the user’s device. An emergency patch shipped on 17 April. The Commission described what it had launched as “still a demo version,” seemingly forgetting it previously described it as ready.

This matters beyond the immediate embarrassment. It is an illustration of the structural argument this post is making. More than 400 researchers from 30 countries had written to the Commission in March calling for a moratorium on deployment until the science on age-verification technology settles. The Commission proceeded anyway. MEP Markéta Gregorová warned the process “is being rushed under political pressure.” MEP Birgit Sippel called it “a half-baked app solution.”

There is no delay in the erosion of anonymity. It begins immediately and applies universally. The question is only how badly the infrastructure built to enforce it fails, and who bears the cost when it does.

A Ban on Curated Content for Minors

Here the logic of the “delay” framing becomes most obviously strained.

The law removes accounts. It does not remove content. A minor who cannot hold an account on TikTok or Instagram still lives in a world saturated with TikTok and Instagram content: in group chats, in reposts, in screenshots, in conversation. The platforms’ content does not stop reaching under-16s because under-16s cannot log in. It reaches them through every other vector the connected world provides.

What the account provided was not exposure to social media content. What the account provided was a frame, a profile, a set of settings, a community, a feedback mechanism, and potentially, a route to help. Algorithmic safeguards, content filters, age-appropriate settings, and reporting tools are all account-level features. Remove the account and you remove the guardrails, not the content.

A minor who encounters distressing content without an account has no report button. A minor who is targeted has no moderation recourse attached to their identity. A minor who finds community has no stable place to return to it. The law, in protecting them from having a social media presence, ensures that their social media exposure is as unstructured and ungoverned as possible.

Von der Leyen’s Copenhagen speech was at its most compelling when it addressed platform design: endless scrolling, autoplay, push notifications, business models that treat children’s attention as a commodity. That analysis is correct. The repository has documented it extensively. But removing the account does not remove the design. It removes the child’s ability to navigate that design with any of the structural protections an account provides.

This is not a delay. It is a specific and consequential subtraction. What has been removed is not access to social media but access to the protections that structured access provides. Describing that subtraction as a pause understates it badly.

A Ban on Having a Voice for Minors

The third ban is the one most consistently absent from policy discussion, and it is the one most directly in tension with von der Leyen’s own framing.

In Copenhagen, she asked: “The question is not whether young people should have access to social media. The question is whether social media should have access to young people.” It is a rhetorically elegant inversion. But it obscures the fact that social media, as it exists, is not merely something that happens to users. It is a speech infrastructure. Billions of people use it to publish, organise, argue, document, and bear witness.

The account is the instrument of speech. Without one, a minor cannot post. They cannot respond. They cannot participate in the discourse that shapes decisions made about them. Under Australia’s law, the cohort most directly affected by the policy is also the cohort that has been formally excluded from commenting on it in the spaces where commentary happens. A European law modelled on Australia’s would replicate that exclusion at continental scale.

The contradiction is sharpest at the European level. In Austria, Germany, Malta, and Belgium, 16-year-olds are already eligible to vote in European Parliament elections. The same EU whose Commission president is now proposing a social media restriction for under-16s counts 16-year-olds among its electorate in four Member States. These young people have been judged capable of choosing the legislature that makes laws for 450 million people. The Commission is simultaneously proposing to remove their access to the primary medium through which political life is conducted, debated, and organised. The logic of democratic participation runs in one direction. The logic of the social media ban runs in the other.

This is not a delay in their speech. It is a prohibition. The question of whether that prohibition is proportionate to the harms the law claims to address is one the repository has tracked extensively. The answer the evidence gives is, at best, uncertain. But the existence of the prohibition itself must first be named accurately before it can be weighed.

Many organisations working on protecting children, such as the 5Rights Foundation, have long argued that children are rights-holders, not objects of protection. Rights-holders have standing to speak. Both the Australian law and the European “delay” concept being discussed in Copenhagen treat the speech of under-16s as a cost-benefit rounding error.

What the Language Is Trying to Obfuscate

Language shapes what gets scrutinised and what does not.

When von der Leyen chose “delay” in Copenhagen, she was probably doing the same thing the Australian eSafety Commissioner has done since December: making a complex set of structural restrictions sound like a considered pause before resumption. It is language designed to reassure. It is also language that makes it harder to ask the necessary questions.

A “delay” is temporary. A ban on anonymity is structural. A “delay” is singular. Three simultaneous restrictions on three different categories of people are not singular. A “delay” implies that what is being held back will be returned intact. An architecture of age verification, once built, does not dismantle itself when a user turns 16.

The eSafety Commissioner chose “delay” because “ban” carries political weight. The word of the year in Australia was “ban” because the population understood what was happening. Europe is now likely being invited to follow the same road with the same map and the same misleading label on the tin.

Three things are being banned. One is temporary. Two are not.

The ban on under-16s having accounts ends at 16. The ban on anonymous participation does not. The ban on content encounters without structural protection does not. And the exclusion of young people from the speech infrastructure through which policy is debated does not reverse itself at a birthday.

Von der Leyen said in Copenhagen that the Commission expects to come forward with a legal proposal this summer. That proposal deserves scrutiny that the word “delay” is designed to deflect. The repository will track it.

Whether any of this is justified is a question the evidence must answer. What it cannot do is answer a question that the framing refuses to ask.

Written by Caroline De Cock, LL.M., Head of Research