The Subjunctive of Images: Alexander Kluge’s Visual Poetics of AI

This paper considers the ontological and aesthetic implications of AI-generated images by looking to the work of the pioneering German filmmaker, philosopher, and writer Alexander Kluge (b. 1932). In recent years, Kluge has taken up images coproduced with the latent text-to-image diffusion model Stable Diffusion in both gallery installations and in his book, Der Konjunktiv der Bilder: Meine virtuelle Kamera (K.I.) (2024). To consider the implications of this seemingly “new” form, we first return to Kluge’s earlier understanding of cinema as rooted in the capacities of the subject and montage. We then examine how Kluge extends his conceptualization of the “virtual camera”—the camera inside our heads, the camera of conditions and possibilities—to his use of Stable Diffusion. Here we accompany Kluge’s methodological self-descriptions with concrete examples, as well as a run-through of the latent diffusion model’s mechanics. Our own method takes on the shape of a montage, as we juxtapose and leap between these different aesthetic and formal operations. This approach allows us, in turn, to consider how Kluge’s retelling of AI’s uses and abuses opens the way for further illuminating readings, including a concluding comparison with the early twentieth-century forefathers of Kulturwissenschaft, Walter Benjamin and Aby Warburg.