Professor in Residence, Department of Architecture, GSD, Harvard University, Cambridge MA, USA
This essay aims to explore the unexpected intersection of AI-generated images, metaphor, and fantasies of knowledge. Expanding on the notions of metaphor in the philosophy of Jose Ortega y Gasset, we view the AI-generated images (such as those from mis-trained StyleGAN models) as ‘objects’ in possession of an inherent surplus that transcends their perceived qualities and constituent parts. Metaphors reveal latent qualities of ‘objects’ that extend beyond human perception. We consider how these image-objects operate not merely as generators of visual metaphors but as metaphor-like entities themselves, shifting our perception of reality. As a result, these image-objects behave like disembodied expressions of a new kind of knowledge. This feeling, or fantasy of knowledge, is what Steven Connor calls ‘epistemopathy.’ By examining the interplay between human imagination, AI-generated outputs, and the epistemopathic experiences they evoke, we explore how AI becomes a manifestation of the human desire for expanded knowledge and imaginative exploration. In the blurring of the boundaries between our own understanding and AI’s perplexing output, we are confronted with a sort of knowledge boundary that conflates reverie with reasoning.