The Reconstruction of the Proto Image

Abstract

Generative Artificial Intelligence, including Generative Adversarial Networks (GANs) and diffusion models, has the potential to significantly influence contemporary art and visual culture. This study examines the unique nature of generative image production through the concept of the "Proto Image," an essential, intermediate visual form that exists independently of physical reality. Acting as a "virtual prototype," the Proto Image emerges dynamically through algorithms, randomness, and symbolic logic, overcoming traditional representation-based logics. It makes the generative process itself a fundamental part of symbolic meaning, revealing the ambiguity and openness of symbolic creation and semantic evolution in AI-generated imagery.
Drawing on Gilles Deleuze's anti-Platonic philosophy and Vilém Flusser’s notion of technical images, the study explores the semiotic and ontological dimensions of generative AI images. By investigating technical elements such as noise and randomness, the Proto Image is presented as a dynamic symbolic network shaped by processes of disruption and optimization. Roland Barthes’ concepts of "Studium" and "Punctum" provide a framework for analyzing the emotional and cultural dimensions of these images and the central role of technology in their interpretation. Interculturally, the study compares the Chinese philosophical concept of "Xiang" (?) with Western semiotics, revealing the Proto Image’s distinctive qualities in dynamic symbolic generation and holistic conceptualization.
The emphasis in "Xiang" on transformation and process resonates with the evolving nature of the Proto Image, reflecting its adaptive and generative dynamics. Although the technological innovations of generative art have been widely discussed, systematic semiotic and philosophical analyses remain scarce. This dissertation establishes an interdisciplinary framework combining semiotics, philosophy, media theory, and fine arts to explore the unique epistemological and ontological status of AI-generated images. It aims to fill a gap in the study of generative art while proposing new theoretical perspectives for its role in contemporary art and visual culture. The study addresses key questions: What is the ontological status of AI-generated images? How do they differ from traditional images? What epistemological potential do they hold, and how do they relate to representation?

Short biography

Chao Li, artist, Phd at the Academy of Fine Arts, Vienna, Austria, Master at the University of the Arts, Berlin, Germany, and at the Academy of Fine Arts, Zurich, Switzerland, Bachelor at the Academy of Fine Arts, Shanxi University, China. Member of the political art group SAN PANG. He is a member of the Austrian Society of Fine Artists, a member of the Berlin Artists' Association, a member of the China Institute of Stage Design, a special artist of the Chinese and Foreign Art Research Institute and a contract artist of the Artinliving Spaces Gallery in Germany.
In his work, Chao Li refers to the socio-political environment and its impact on society. He uses various media such as painting, sculpture, installation, video and photography to explore the relationship between architecture, gender roles, control and power. He often includes the viewer as well as the exhibition space and its surroundings in his reflections.