Against Medium Essentialism: AI-Generated Images and the Reconfiguration of Medium Classification under the Post-Medium Condition

Abstract

【Abstract】This article situates AI-generated images within the theoretical context of the post-medium condition and examines how they reconfigure the classificatory logic of medium essentialism. It argues that the question of medium should not remain at the level of deciding whether an image belongs to a given medium, but should turn instead to the relations among generative mechanism, visual properties, functional operation, and institutional recognition. On this basis, the article advances a new logic of medial-property configuration. Under the post-medium condition, the multiple dimensions of the image have not disappeared; what has been loosened is the necessary connection once guaranteed by medium ontology. The theoretical force of AI-generated images lies in their capacity to detach older medial properties, such as the photographic, the pictorial, and evidentiary force, from their original material mechanisms, and to allow them to return as forms of institutional memory. In this way, AI-generated images disclose the bound relations of properties that have long been embedded in traditional medium classification.