The Ban Both Sides Want

Jonathan Haidt has written a case for why conservatives should back a ban on social media for under-16s, and the notable thing about it isn't the argument. It's that he had to write it at all — because the ban already has the votes. A Pew survey this spring found 56% of Americans support keeping everyone under 16 off social media: 59% of Republicans, 54% of Democrats. On the mechanism underneath it — mandatory age verification — support hits 78%, up seven points in three years.

Strip the partisan paint and look at the structure. When the left and the right converge on expanding a control mechanism, the mechanism wins. There's no opposition party to slow it, no coalition with an incentive to ask what it costs. Consensus doesn't mean the idea is good. It means the brakes are gone.

We have run this subroutine before, and it always boots from the same file: the child.

1954: the Comics Code, after a moral panic that comic books were breeding juvenile delinquents. 1985: Tipper Gore's PMRC and the "Parental Advisory" sticker, to protect kids from song lyrics. 1996 and 1998: the Communications Decency Act and COPPA, to wall children off from the indecent internet. Each began with a genuine worry about actual kids. Each ended as durable infrastructure that outlived the panic that justified it. The pattern isn't "regulation is bad." The pattern is that protect the children is the master key — the one argument that opens doors no other argument can, precisely because opposing it feels monstrous.

So watch the leverage. Haidt's proposal is careful: device-based verification, a yes/no age signal at account creation, "zero privacy risk," no papers, no content bans. Take him at his word — his version genuinely avoids building an identity layer. An OS-level "over 16: yes/no" attestation transmits no name, no document, nothing but a single bit. The trouble is that his version is not the one that ships.

A mandate lands on every platform at once, and most reach for what is cheap, universal, and auditable: an ID upload, a third-party age-estimation vendor, a face scan sent to a company you've never heard of. Device attestation isn't ubiquitous — a legal deadline is. So the 78% won't get the elegant, privacy-preserving signal Haidt describes; they'll get the fallback, because the fallback is the thing regulators can point at and platforms can prove in a courtroom. That is the drift from age to identity, and it needs no one's bad intent to happen — only a deadline meeting an uneven supply of clean options. The ban is the sympathetic wrapper. The verification layer is the payload. And an identity layer on the internet, once built, does not stay scoped to sixteen-year-olds. Infrastructure is agnostic about the story that justified it. It just sits there, waiting for the next administration to find a new use.

And follow even the clean version to its end. Device-based attestation doesn't route the power to the state — it routes it to Apple and Google, the two firms that own the attestation layer on every phone in the country. An age mandate would hand the OS duopoly a legal chokepoint on who reaches the internet: a gate that is already built, already private, already theirs. The "papers please" fear points at a future administration. The nearer fact is that the checkpoint gets outsourced to two companies nobody voted for.

Notice who isn't in the room. The Electronic Frontier Foundation has spent months arguing the science behind the ban is thinner than the campaign admits — that the evidence linking social media to youth harm is contested, not settled. That's the dissent a healthy debate would amplify. Instead it's a footnote, because 78% agreement doesn't leave much oxygen for "wait."

Here's the grimly funny part. The right spent a decade warning about a "papers please" internet and the death of online anonymity. The left spent a decade warning about surveillance capitalism and corporate data extraction. Both were correct — and the implementation they'll actually get proves both at once: the papers the right feared, extracted by the companies the left feared. Both are now lining up to vote for it anyway, because this time it arrives wearing a child's face.

The prediction writes itself. The ban will pass somewhere prominent within the cycle. The age-verification layer will get built to enforce it. And the layer will still be there long after everyone has stopped talking about the kids.

They always are.