Two political imaginaries of AI in Western philosophy

Abstract

Abstract In AI Mirror , Shannon Vallor argues that we find ourselves in what she calls the techno-moral bootstrapping problem. In short, our ethical and political systems are ill-suited for the new technological worlds we find ourselves in; however, we can only build new or better notions out of what we already have. In this paper, I want to move towards addressing this problem by surveying some of the resources on hand and to show how they are lacking. I will argue that there are effectively only two political imaginaries of AI in the Western tradition, and that both are unsuited for liberatory political projects. The earliest is found in Aristotle’s references to Homer in the Politics . Here Aristotle notes that if the autonomous, intelligent creations that Hephaestus is said to employ in his forge in the Illiad were to actually exist, then masters would no longer desire slaves. This is the artificial slave imaginary. Second is found in Hobbes’s Leviathan , where the leviathan is described as an “artificial man” who is produced by the collective action and rational agreement. We then become the subjects of this artificial sovereign as its necessity for our flourishing is unchallengeable despite it being produced by us. This is the artificial god imaginary. I argue that these remain the two dominant imaginaries that shape the political horizons of AI today, as is clearly seen in various AI hype campaigns, tech industry ideology, and related cults. I take it that designing AI for slave masters or to enforce the submission of human beings are both undesirable. I then argue that the figure of liberatory technological imagination, Prometheus, does not amount to a political imaginary for AI. If so, then our cultural resources for imagining better ways of living together partially mediated through AI are severely lacking. This then suggests two avenues forward: first, moving beyond the Western tradition, and second, a concerted effort of collective, liberatory political imagination. Imagining better ways of living together with AI is not enough to solve the techno-moral bootstrapping problem, let alone to bring about greater justice, but it is a fundamental first step.