Scott Allen
Ambiguous Boundaries -windows-
Ambiguous Boundaries -windows- is an interactive installation in which a display mounted on the gallery wall functions as a window. On screen, pre-recorded footage of the space just outside the gallery plays continuously—an ordinary, unremarkable scene. But something is wrong: the people peering in through this "window" wear the viewer's own face. Their clothes are different, their hair is different, but the face is unmistakably yours.
At the start, the viewer's face is photographed. A deep learning-based face swap system then transplants their features—in real time—onto figures in the pre-recorded video. The result: strangers outside the window are looking in at you, wearing your face.
Deepfake technology is conventionally used so that one person can assume another's identity—I become someone else. This work inverts that operation entirely: here, others become you. This reversal transforms deepfake from a tool of self-projection into a mechanism of involuntary self-dispersion—your face escapes you, proliferating across bodies you do not control.
The "window" frame is central to the work's meaning. A window is where inside meets outside, where your private space ends and the public world begins. In this installation, that spatial boundary mirrors a psychological one: the boundary between self and other. When the outside world is populated by your own face, both boundaries dissolve simultaneously. The viewer's sense of bounded selfhood is destabilized—not by becoming someone else, but by discovering that someone else has already become them.
By inverting the directionality of deepfake, the work raises a question not about the reliability of images, but about the stability of selfhood—opening a space for reflection on identity, authenticity, and the increasingly porous boundary between the self and its computational doubles.
Completed IAMAS in 2016. Focusing on the relationship between human imagination, visual devices, and technology, he creates installations and performances by intervening in projection mechanisms and reworking everyday objects into images. He makes deep-learning art and performs as Ai.step, an AI live coding unit. Awards: Best Works Award at CVPR 2024 AI Art Gallery, Digital Choc 2019 Grand Prize, Yamanashi Media Arts 2021 Excellence Prize. Festivals: FILE 2025, MUTEK Montréal, Scopitone 2019.




