Generative and Adversarial: Art and the Prospects of AI

Abstract This essay serves as an introduction to the Questionnaire on Art and Machine Learning. It opens with the testimony of Sam Altman before the US Senate Judiciary Committee on Privacy, Technology and the Law in May 2023, following the launch of ChatGPT a year earlier. ChatGPT is part of a wave of generative AI: AI that not only classifies or analyzes information but also generates it. Media responses to ChatGPT and a range of other text-to-image and text-to-video generators—including Dall-E, Stable Diffusion, MidJourney, Deep AI, and Sora—tend to divide between “Boosterism” and “Doomerism”: They either echo Silicon Valley's habitual refrain about the inherent goodness of technological innovation or they describe dystopian scenarios in which human agency is all but eclipsed. “Generative and Adversarial: Art and the Prospects of AI” argues that artists, art historians, critics, and curators offer far more nuanced perspectives on the prospects of AI. Borrowing the language of “generative adversarial networks”—one of several machine-learning models that have revolutionized image production in the past decade—it interrogates AI's aesthetic, political, and ethical possibilities, as well as its historical genealogies.