Nobody knocked.
The machine let itself in anyway.
On the day an AI feature enrolled every public account on a platform with nobody’s consent — and why recognition and consent turn out to be the same failure, run in opposite directions.
Nobody knocked. On July 7, Meta launched an AI image generator built into Instagram — tag any public account’s username in a prompt, and it pulls from that person’s public photos to generate new images of them. Every public account was enrolled the moment the feature shipped. No screen, no prompt, no moment where the person attached to that account weighed in before a stranger could generate an image of them. They found out the way everyone eventually finds out about a decision made in their name: after it had already happened.
This is not a story about a bad toggle. It’s a story about a door that was never built. Consent requires a moment — a screen, a pause, a question a person is free to decline. Remove the moment and you don’t get less consent. You get the appearance of consent with none of its substance: an enrollment logged as agreement, because no one designed the point at which disagreement was possible.
Recognition and consent are the same failure, run in opposite directions.
Digital Derangement Syndrome™ describes a machine that fails to notice you exist — the recognition never forms, the citation never lands, thirty years of real authority goes unrecommended by the system deciding who to surface. What happened here is the mirror image of that failure. The machine noticed the person exists, and acted on it, without ever bringing them into the room. This is Trust Transfer Failure running in the opposite current: instead of expertise failing to translate into machine-readable trust, a person’s likeness and data translate into machine action without the trust ever being asked for. One version of the Inference Era ignores you. The other decides for you. Both share the same missing element: a person, present, at the moment the system made its determination.
Default flipped. The feature shipped opt-out, not opt-in — running by default is a design decision, not a technical necessity.
Notice omitted. No screen, no summary, no disclosed moment surfaced the change to the person it affected before it acted.
Silence logged as agreement. The absence of an objection became, functionally, a recorded yes — the system had nothing else on file, so it treated not-stopping-it as consent.
The correction is the same discipline, whichever direction the failure runs.
Answer Engine Authority™ exists because a system’s assumptions about who you are get made with or without your input — the only real choice is whether you designed them. The same discipline that installs a corrected signal so an engine finds your authority before it invents a gap applies exactly when the failure runs the other way: nothing forces silence to become load-bearing. It is simply cheaper than building the door.
The feature lasted three days. Backlash forced Meta to remove it on July 10 — not because consent had been designed in before launch, but because enough people noticed after the fact. The door still wasn’t there when the feature shipped. It only got boarded up once enough people had already walked through the one that wasn’t.
Fourteen editions into this record, every one of them has turned on the same clock: something about you — your authority, your fact pattern, your consent — gets decided by a system that never checks whether you’re in the room. What happened on Instagram is not a new failure. It is one this record has already been describing, run at scale, in daylight, with a byline — and a three-day shelf life once the daylight arrived.
Meta launched an AI image generator (“Muse Image”) inside Instagram on July 7, 2026, automatically enrolling all public accounts with no user notification; the company removed the feature July 10 following backlash — reported by CyberInsider, Digital Trends, and Tom’s Guide.