The world screened: Generative AI and us

Abstract

For many commentators, user-friendly generative AI technologies, such as Dall-E, ChatGPT, and Runway, herald an onslaught of unreal, informatic simulacra that eclipse the creative work of reading, writing, watching, and making. Media theorist Matthew Kirschenbaum, for example, has predicted a “textpocalypse” in which human-authored texts will be lost in a sea of machine-generated facsimiles. A flashpoint in these debates was the 2023 Hollywood writers' strike, which sought to limit the corporate exploitation of AI as a tool for replacing human workers. But is the situation really so simple and well-defined, such that we can speak of an inhuman informatics that stands opposite to the supposed agency, originality, and critical spirit of human readers and writers?

This talk argues for a new formulation of the problem, attentive to how recent AI technologies tend to catalyze, or even embody, anachronistic humanisms. Among other shortcomings, this simplistic binary creates to more or less purified poles—human and machine—with little or no “mediation” operating between them. Only by transfiguring the entrenched opposition between technics and the human can we achieve a properly humanist (or, for that matter, anti-humanist) philosophy equal to the demands of the present. In particular, I offer a sketch of a critical theory that privileges “intermediation” over opposition in accounting for the work of emerging AI technologies.