On artificial intelligence, crisis, and governmentality

Abstract

This thesis offers a Foucauldian perspective on AI governance by exploring the existential crisis from superintelligence as a governmentality issue with both biopolitical and geographical implications. By drawing insights from the current literature on the existential risk from AI as well as biopolitical and geographical works on Foucault’s concept of governmentality, this thesis argues that while the conceptualization of intelligence within the context the existential crisis suggests a novel kind of biopolitics which conceives humanity in terms of cognitive capabilities, the spatiality of the notion of AI highlights the need for a geography of governmentality which accounts for how such notion enables the formation of ensembles of global catastrophic risks according to particular rationalities of governance.