Leonardo Impett
There Is No Generative AI: Epistemic Compression and Neural Exchange Value 

This talk draws on a central thesis from a forthcoming book, Vector Media (Impett, Offert, Drucker). Vector Media traces the historical epistemology of artificial intelligence, and highlights the latent theories of the image, and of knowledge, inscribed in deep neural networks. The inevitable “bias” of these systems lies not only in what they represent, but in the very logic underlying representation. Their internal ideologies are often not visible in training data or generated outputs, but emerge from how the model organizes and transforms information within itself.  

While previous media technologies have created new formats or imitated existing media, neural networks instead seek to dissolve all previous media into a universal space of commensurability: vector space. These models impose operational commensurability on media objects — from representation to interpolation, from medium to vector — making possible their circulation, translation, and manipulation within neural infrastructures. Vector Media thus shows how AI's internal mechanisms inscribe new epistemologies of media. Cultural objects, once specific to a medium, are now rendered fungible — commodities in a new neural economy, expressed solely in terms of their "neural exchange value." 

BIO
Leonardo Impett is Research Group Leader at the Bibliotheca Hertziana, Assistant Professor in Digital Humanities at the Faculty of English, University of Cambridge, and Bye-Fellow of Selwyn College. His research focuses on crossing AI and visual culture in both directions: exploring AI methods in digital art history, and in examining bias and visual ideology in AI from the perspective of art history and visual studies. He frequently works with machine learning in arts and culture, including with the Liverpool Biennial, the Royal Opera House, and the Whitney Museum of American Art. 

Leonardo has a background in information engineering and machine learning, having worked or studied at the Cambridge Machine Learning Lab, the Cambridge Computer Lab’s Rainbow Group, Microsoft Research, and EPFL’s Image and Visual Representation Lab. 

Æstetisk Seminar er tilrettelagt af AIIM – Centre for Aesthetics of AI Images, Lotte Philipsen og Mette-Marie Zacher Sørensen, Institut for Kommunikation og Kultur, Aarhus Universitet.