2025/08/01 by Lindberg, Susanna
#Artificial intelligence #Derrida #Imagination #Kant #Mimesis #Technology
paper · doi:10.15160/2282-5460/3030
This article reflects on the philosophical presuppositions of artworks made with the help of AI. Such works illustrate the author’s more general claim that in order to understand works-of-AI, we should not look at them as works of intelligence (aiming at logos) but at most as works of imagination (playing with mimesis). But can it even be said that AI imagines? To clarify this issue, the article follows the idea of imagination as formulated by Kant and then developed by Derrida, Lacoue-Labarthe and Nancy. In this line of thought, it becomes possible to think that imagination is not only a free gift of nature, but always already a technics. For these philosophers, it is a particular technics whose law is mimesis, that is, capacity of producing and sharing semblances, fictions and figures. Is this also what a generative AI does? The article distinguishes more and less imaginative uses of AI in art and suggests, in the end, that even if it cannot be said that a generative AI imagines, still, in best cases, imagination can take place in well-constituted assemblies between humans and AIs.