Wor(l)d-Image Trans-formation: Looking through DALL-E 2 and Midjourney

Abstract

Walter Benjamin pondered the relationship between image and caption, wondering whether the caption would become the most meaningful element of an image. But what if the caption becomes the image? The relationship between caption and image has been opened to new exploration by recent Artificial Intelligence (AI) text-to-image generation models, such as DALL-E 2 and Midjourney. Models such as DALL-E 2 and Midjourney generate images when the user submits a written prompt, as little as a single word or a phrase. Following Yuk Hui’s concept of cosmotechnics and Joanna Zylinska’s post-humanist paradigm, this paper analyzes DALL-E 2 and Midjourney as vision technologies offering both an opening and closure to the Greek notion of techné in the twentieth-first century. First, DALL-E 2 and Midjourney are discussed as algorithmic generators of an already established visual dictionary. The algorithmic image generation is thus criticized as reconfiguring human imagination through a double-rationalization process (language – calculation – image) and threatening human sensibility. Second, DALL-E 2 and Midjourney are discussed as non-human entities capable of surpassing human ways of seeing. Following the intrinsic relation between art and technology, the second part of this paper focuses on if and how a new sensibility of the ‘outside’ can be achieved through technology such as Midjourney and DALL-E. The possibilities of such models are argued to carry the possibility of allowing for further visual engagement between the human and non-human approach to visuality.