Frags is a series within Native, a research project in which I sow specific visual conditions to invite AI image-to-video systems, unprompted, to reveal their own perceptual logic—then harvest and classify the resulting gestures.
Each Frag is a looping, single-object video: a kind of digital sculpture that captures the uncanny moment when AI simulates rotation in space. As the object turns, the machine fabricates surfaces absent from the keyframe image and calculates how light and color might refract or reflect across them. The behavior resembles an optics engine inventing its own physics more than it does traditional animation.
Like the early “Cinema of Attractions,” Frags hold attention through pure spectacle, reminding us that wonder is part of cinema’s original DNA. They also extend digital video’s own fascination with luminous display—backlit, looping, and made from emitted rather than reflected light, a medium where spectacle and luminosity converge.
The imagery across Native and Frags draws from both documentary and New Topographics traditions—from unremarkable built environments to those marked by conflict, collapse, or neglect—reframed not as spectacle but as terrain where the machine tests its idea of reality, offering us an uneasy mirror.