Signal, Noise and Complexity as Poetic Vectors in Recent Generative Videoart

Abstract

This article surveys recent videoart practices which employ generative artificial intelligence (AI) and machine learning resources, distinguishing between three poetic vectors (currently in evidence in the field: creative directions concentrating on aspects of signal, noise and complexity in such systems. These notions are proposed through the analysis of works of interest and the past repertoire of videoart practices with which they sustain ties, also answering to the modus operandi of AI models and questions relevant to “computational art”, as addressed by Abraham Moles, among others. A contrastive debate between these creative directions is elaborated, focusing on works by Refik Anadol, Casey Reas and Pierre Huyghe, as a way of highlighting distinctive characteristics of their recourse to AI models, mapping current artistic directions in exploration. On the final remarks, the role of artists as social mediators of these technologies is reflected upon.