When the Medium Fabricates the Experience: AI-Generated Video and the Representation–Experience Boundary — Part VI of a Research Program on Communicative Compensation

This research program argued that video-mediated communication (VMC) — Zoom calls, YouTube tutorials, WhatsApp video messages — is adopted as a compensatory adaptation that restores part of the naturalness suppressed by leaner channels, provided a real interlocutor or a real recorded event is on the other end of the exchange. This paper, Part VI, follows the same channel to the point where that condition no longer holds: video that is not a recording of anything and connects to no real interlocutor at all, but is generated in its entirety by artificial intelligence. A video adds what an image structurally cannot supply: duration, motion, rhythm, and causal sequence — the raw material from which human perception, according to event segmentation theory, constructs its basic units of understood experience. When video is generated entirely by artificial intelligence, the product becomes a candidate for what presence theory calls an illusion of non-mediation. Because human reality monitoring relies heavily on exactly the perceptual and contextual richness that generated video can now supply, this paper argues that fully AI-generated video raises a different order of question: not whether a representational skill is compensated for, but whether a simulated sequence can substitute for the having of an experience itself. Reviewing recent empirical work on AI-edited and deepfake video and false memory, this paper proposes a Representation–Experience Boundary framework distinguishing structural components (temporal structure, perceptual presence, and source attribution) that jointly determine whether a generated video functions as a labeled representation or as unlabeled evidence of an event.