Why do we believe a shaky dashcam or doorbell-cam clip before we even ask where it came from?

This explores why footage that *looks* like raw evidence, such as a security-camera angle, a dashcam timestamp or an unedited documentary shot, gets believed before anyone asks where it came from. The corpus doesn't study camera framing or film theory directly. It does have a sharp set of notes on the underlying mechanism: belief gets triggered by the *form* of evidence, not by evidence itself, and that form can now be copied cheaply.

The most direct doorway is Horning's argument about AI slop Does AI-generated slop exploit visual truth to bypass skepticism?. Generative video models have absorbed the visual tropes of evidence from billions of images: the static, slightly-too-high mounted angle, the grain, the uncomposed framing. They reproduce those tropes on demand. The point isn't that viewers are fooled in the old sense. They get the *feeling* of having witnessed something real, with no obligation to check it, believe it firmly or act on it. Evidence turns into content, and the mounted-camera look becomes a style that signals "nobody staged this" even when everything was staged.

Why does the style work so well? A linguistics finding offers a surprising parallel. Presuppositions persuade better than direct assertions Why are presuppositions more persuasive than direct assertions?. "Even the mayor stopped lying" slips in a claim as already-settled background, so it skips the scrutiny an outright statement would get. Documentary framing does the same thing visually. A fixed, unattended camera doesn't *claim* "this happened." It presupposes it, because the framing implies no author and no argument, just a recording. Since nothing is being asserted, there seems to be nothing to push back on.

The experimental evidence suggests people have no fallback once that cue is faked. In an 81-person study, readers given no provenance signals couldn't tell truth from fluent fabrication at all, and performed at chance Can readers tell truth from fabrication without evidence signals?. Discernment came back only when an interface showed which claims were actually verified. The lesson for video: our sense that something "looks real" was never doing the verifying. It was standing in for a provenance chain we assumed existed behind the image.

That ties into a bigger reframing: AI output works like pre-Enlightenment hearsay, a retelling with no traceable origin that can't be checked against a stable source Does AI-generated knowledge have the same structure as hearsay?. Documentary indexicality, meaning the idea that a photo is physically caused by what it shows, was one of the strongest tools against hearsay. Synthetic footage keeps the look of that causal link while cutting the link itself. So the mounted camera can trigger belief without context because the context it used to carry inside it is exactly what has been lost.