The Trust Recession
It could be her voice — or it could be a fake.
The question
My phone rings at 11 at night and it's my sister's voice, and she's crying, and she says she's in trouble and she needs money sent right now and please don't tell Mum. And for one second my whole body believes her — I'm already reaching for my wallet — and then the other thing arrives, the thing that lives in all of us now, and it says: that could not be her. That could be three seconds of her laugh scraped off a birthday video and fed to a machine. So I do the thing you're supposed to do. I ask her something only she would know. And she answers it. And here's what I can't get out of my head afterward — it didn't help. Because I sat there thinking, a good enough fake would know that too. There is no question I can ask that a good enough forgery couldn't pass, and I knew it, and I sent the money anyway, shaking, not because I'd proven it was her but because I'd run out of ways not to.
And it didn't stop at the phone call. The next morning I caught myself doubting a news clip I'd have believed without blinking a year ago — real footage, as it turned out — for no reason except that now everything wears the same face. I used to know things by seeing them. Now I've started to disbelieve the true ones, just to be safe. I thought the problem was that I might get fooled. The problem is worse. I've stopped being able to be sure of anything I didn't personally watch happen — which, I'm realising, is almost everything I have ever claimed to know.
The episodes
Each episode builds on the one before, and the finale hands the question back to you sharper, not settled.
8 of 8 transcripts are readable now — audio is coming.
-
1
It could be her, it could be a fake
A late-night phone call in a sister's crying voice, and the realisation that no security question could rule out a good enough fake. The first episode offers no test for detecting forgeries — there is none — and instead uncovers the assumption the panic rests on, that real knowledge is what you check for yourself.
Read transcript → -
2
Proportion your belief to the evidence
David Hume's rule — a wise man proportions his belief to the evidence, and another person's word is evidence only once you have grounds that they are reliable. The episode builds the case for the verify-first instinct as strongly as it can, and holds one question back for the end — not whether you should live this way, but whether anyone can.
Read transcript → -
3
Reid's two principles
Thomas Reid's reply to Hume — nature built us to believe what we are told, and testimony is a basic source of knowledge, like sight, owed belief until something specific breaks it. Hume says trust must be earned; Reid says we start out already owing it. Both pictures of the human knower are left standing, with notice that you cannot keep both.
Read transcript → -
4
Name one thing you know by yourself
An inventory rather than a lecture. The listener is asked to name one important thing they know with nobody's word in it — a birthday, what is in the pill, that the distant war is happening — and watches each dissolve into testimony. One reply is held over for next time — grant that checking everything is impossible, but vet your sources, then trust them.
Read transcript → -
5
You can't check the checkers
C. A. J. Coady closes the last exit — every tool you would use to check a source's reliability was itself built out of trusted testimony, so there is no testimony-free place from which to run the audit. Trust turns out to be not a lazy shortcut but the ground, and the question becomes whom to trust, and how well.
Read transcript → -
6
When someone tells you
Richard Moran's account of assurance — a speaker does not leave evidence lying about like a footprint, but stakes themselves on what they say — and Annette Baier's account of trust as accepted vulnerability. Together they explain why a lie from someone trusted is a betrayal rather than a mere false belief, and what a forged voice counterfeits and cannot supply.
Read transcript → -
7
When recordings stop settling it
For two centuries a recording could catch a liar. Don Fallis and Regina Rini argue that deepfakes remove that backstop, so the deepest harm is not the lies we will believe but the true things we will no longer be able to believe; Joshua Habgood-Coote replies that the backstop never bore that weight. The suspicion is tested rather than settled.
Read transcript → -
8
The forced bet
The final episode brings the four accounts into one room — Hume's verifier, Reid's default trust, Moran's staked assurance, and the failing backstop — and crowns none of them. No test for fakes is offered; what remains is the question of what makes a bet on a person a good one, and what is traded away by refusing to bet at all.
Read transcript →
Thinkers
The thinkers and traditions this series draws on: