Abstract

Recent narratives and debates surrounding long-term AI concerns—the prospect of artificial general intelligence in particular—are fraught with hidden assumptions, priorities, and values. This paper employs a humanistic, narrative-centered approach to analyze the works of two vocal, and opposing, thinkers in the field—Luciano Floridi and Nick Bostrom—to ask how the representational, descriptive differences in their works reveal the high stakes of narrative choices for how we form ideas about humanity, urgency, risk, harm, and possibility in relation to AI. This paper closely reads Floridi and Bostrom using different representational models and historical narratives from works in the environmental humanities, literary theory, bioethics, and the history of technology to uncover the imaginative terrain of recent long-term AI discourse and reveal the complexity and limitations of the messaging underlying the works of different authors.