The Age Gate

A law arrives wearing the one frame nobody can argue with: protect the children. It bans social media for everyone under sixteen. But to keep a fifteen-year-old off a platform, you must check the age of every adult who uses it — and the only way to check an age is to verify an identity. So a measure about children becomes a measure about everyone's papers. Watch what it actually builds, and what it leaves untouched: the product is never made safer, the independent watchdogs are folded into a single regulator that polices the speech and holds the data at once, and within weeks the children themselves have routed around it — while the identity check stays, for good, on everyone else. In Canada, the bill is carried by the Minister of Identity.

A note on what this is, and is not

§01 — Four moves, one gate

The same discharge moves that run beneath every case in this series — the Grammar of the Con — run here too. Redefine the standard (ban access, never fix the product). Let the gate stand as the safety. Delete the independent oversight and gate the complaint. Then install the apparatus and let it outlive the reason. Four countries, one technique; read the last column, where the stated goal meets what was actually built.

§02 — Move one: ban the access, never fix the product

A product engineered to maximise engagement, to extract data, to keep a user scrolling past the point they meant to stop — that is the thing causing harm. The age gate does not touch it. It moves the standard sideways: not make the product safe, but keep the wrong users out. The design is never examined; only the door is.

Age limits "won't fix what is wrong with online platforms."Mozilla · on Australia's ban · Dec 2025

Canada makes the inversion explicit: under the Digital Safety Act, a platform that meets safety standards can obtain an exemption from the ban. Safety is the off-ramp, not the requirement. The default is exclusion; the reward for a safe product is merely to be left alone.

Counter: name the standard the product must meet — not the users it must exclude. A safe product is safe for the people on it.

§03 — Move two: the gate becomes the safety

Once a platform can show it took "reasonable steps" to check ages — facial estimation from a selfie, an uploaded government ID, a linked bank account, an inference drawn from your behaviour — its duty is discharged. The verification step becomes the safety step. Australia measured success not by any change in the product but by exclusion: by mid-December 2025, platforms reported removing 4.7 million under-16 accounts. An exclusion count, presented as a safety result. The harm the product does to the people still on it is exactly where it was.

Counter: the gate ran — and the product is unchanged. That gap is the tell. This is the move The Certifiers documents in another domain: the check stands in for the thing checked.

§04 — Move three: delete the watchdog, gate the complaint

The first draft of this Edition guessed the tell was a carve-out — that the ban would skip the newest risk to children, the AI chatbots. The tabled bill closed that guess. Bill C-34 names chatbots as a regulated class and writes them conduct rules: no posing as a human, no impersonating a doctor or a lawyer, no encouraging self-harm or suicide (§ 53). We flagged the carve-out as unverified; the text shut it; we follow the text. The real tell is sharper, and it is a consolidation. In the same bill, the independent Digital Safety Ombudsperson — the office whose whole job was to stand on the public's side — is removed, its role folded into the single Commission that also writes the rules, runs the audits, and levies the fines. One legal reading calls the result an "internet super-regulator." And reaching it is gated: under § 68(2), you are "not entitled to make a complaint to the Commission" until you have first exhausted the operator's own internal process. The offender's process becomes the filter on the offender's accountability.

Counter: an oversight body you must exhaust the offender to reach — beside a watchdog that has been deleted — is not oversight. This is the move Case 10 · The Process Is the Filter documents.

§05 — Move four: the apparatus that outlives the reason

Here is the hinge the whole thing turns on. You cannot confirm that a user is not a child without checking that user — which means checking everyone. A measure aimed at a few fifteen-year-olds requires that every adult prove who they are to read, post, or speak. The age gate is, structurally, an identity gate.

Age-verification systems are, at their core, surveillance systems.Electronic Frontier Foundation · Dec 2025

What gets built is a verification layer over the open internet: ID scans, face databases, a centralised and breachable trove of who-is-who, anonymity eroded by default, a system the EFF notes is "ripe for abuse in countries with weak rule of law." Roughly half of US states are already down this road. The frame — the children — is temporary. The apparatus is not. This is the ratchet Case 23 documents: legibility, once installed, becomes data-power and does not retract.

And the apparatus has somewhere to go. Five days after C-34, a second bill — Bill C-36, the privacy bill — removed the Privacy Commissioner's authority over the private sector entirely and handed it to the same Commission, renamed the Digital Safety and Data Protection Commission of Canada. So the office that decides what speech comes down is now also the office that governs the identity data this very law compels every adult to surrender — and the two independent watchdogs that stood beside those functions, the Ombudsperson and the private-sector Privacy Commissioner, are both gone, deleted in a single fortnight. One office, both mandates, no guard. In Canada, fittingly, the bill is carried by the Minister of Identity. The fuller account is Case 52 · The Merged Office.

Counter: ask what stays after the rationale fades. A protection that builds permanent identification — and merges the body that holds it with the body that polices speech — is not, finally, a protection.

§06 — The children leave; the papers stay

The cleanest disproof of the child-safety story is behavioural, and it is already in. When the UK switched on age verification in July 2025, VPN sign-ups jumped 1,400 to 1,800 per cent — Proton called it "levels usually associated with civil unrest"; a repeal petition gathered close to 500,000 signatures. And Childnet found that VPN use rose among children in the months after — "the precise population the legislation was designed to protect." In Australia, under-16s documented their workarounds openly; some passed the face-scans with game screenshots and AI-generated faces.

So the children leave in an afternoon, and the identity check stays on everyone else — permanently, and heaviest on those with the least margin: the 43% of transgender people without ID matching their name, forced to out themselves or log off; the racialised users whom age-estimation algorithms read wrong; the whistleblower, the abuse survivor, the dissident, for whom anonymity was the point. The thing that was supposed to protect the vulnerable is the thing that exposes them.

It was never about the children. It was about the papers.