Resumo

This paper examines the creative process behind the conceptualization of a video composed of AIgenerated images, structured around a digitized version of a classic studio portrait captured with an analog large-format camera, which is progressively deconstructed and reconfigured into an abstraction.
The research finds grounding in artworks that employ transformation and disappearance as a means of generating new visual forms, whether through mechanical or technological interventions, by artists such as Robert Rauschenberg, Wolf Vostell, Stan Brakhage, Hiroshi Sugimoto, and Thomas Ruff, whose practices—despite differences in medium, methodology, and historical context—navigate the paradox of destruction as creation.
The use of generative-AI in the project aligns with a lineage of experimental approaches to abstraction in video art. The image-making process, rooted in human-machine collaboration, unfolds through an iterative approach in which the creative vision of the author is negotiated with the statistical mechanisms of AI-driven synthesis. A dual-input strategy is employed by combining image and text prompts, with the latter incorporating terms related to unseen layers of digital technology, based on the premise that subjectivity can foster abstraction.
Situated within broader discussions on the evolving relationship between technology and visual culture, this study interrogates the conditions defining the current landscape of abstract moving-image production. It examines the transition to probabilistic imaging as a shift that expands abstraction beyond aesthetic concerns, framing it as both a form of conceptual invisibility and a critical mode of inquiry into underlying mechanisms of contemporary digital imagery.