Presumed Synthetic
The benefits line has a verified channel that settles claims in seconds, and a queue for everyone else. June has been on hold for forty minutes, being asked, for the third time today, to prove she is a person.
This story is from the AI Scenario Explorer, which pairs eight forces shaping AI, two at a time, into 28 maps — each map with four corners, each corner a different way that pairing could play out. A vignette takes one corner and makes it concrete: structured conjecture, not prediction. About the project →
The queue for humans is longer now.
June has been on hold for forty minutes, listening to a recorded voice explain that her call is important and unusual. Most of the department's business arrives through the verified channel: accredited agents, speaking machine-to-machine, settle ordinary claims in seconds. The phone line is what remains — the fraudsters, the confused, and June, who is neither.
She is renewing her mother's care support. Last year this was a form. This year the form sits behind a sign-in that wants proof she is a person, because last year the department received four million applications and found that most of them had never breathed. Nobody chose cruelty. They chose triage. When anything that can fill in a box can claim a benefit, an open box is an open wound, and every check that follows is somebody's reasonable fix.
Her phone is eight years old. The app that vouches for people will not run on it. The fallback is a video check at the library: turn your head slowly to the left. The strip light flickers, the check fails, the queue behind her watches her fail it again.
Her nephew has never seen any of this. His identity settles silently — his bank app, his handset, six years of contactless breakfasts vouch for him before each page finishes loading. The system does not distrust June because of anything she did. Her file is thin: no passport, rented all her life, a prepaid phone, cash where possible. The evidence of being real turns out to be a purchase history, and she never bought enough of it.
The deep loop takes her a week to see clearly. The personhood app wants a bank account in good standing. The bank, since the flood, wants the personhood app. Each requirement is reasonable. The circle they form is not — and the circle is invisible from inside either office, which is why no one who built it has ever stood in it.
It is the boy at the library who finally solves her, kindly and correctly. There is an assistant now, £4.99 a month, and the cheap tier is enough: it holds a platform-issued credential, insured, whitelisted, machine-readable. It files the renewal in eleven seconds. The department believes her agent in a way it could not afford to believe her.
She is the only thing in today's queue with a pulse, and she got through it by sending the machine ahead to knock.
This is a vignette from the AI Scenario Explorer — structured conjecture, not reportage: one corner of Access × Epistemics given room to breathe. Signs it is arriving: personhood checks spreading across public and commercial services as agent traffic grows; device-attestation requirements that quietly exclude old or cheap hardware; identity systems that fail people with thin credit files; paid verification that clears in seconds beside free queues that clear in weeks. Signs it is not: a statutory right to a human channel at equal service speed; free public proof of personhood that works offline and on old phones; verification burdens placed on institutions rather than claimants; fallback checks designed around the people most likely to need them.
The checks exist because the flood is real. The question a society gets to decide is who pays the toll of disbelief — and the quiet default, one reasonable fix at a time, is that suspicion becomes a tax on being poor, while being believed becomes something you subscribe to.