Human-centric AI: philosophical and community-centric considerations

Abstract

This article provides a course of correction in the discourse surrounding human-centric AI by elucidating the philosophical underpinning that serves to create a view that AI is divorced from human-centric values. Next, we espouse the need to explicitly designate stakeholder- or community-centric values which are needed to resolve the issue of alignment. To achieve this, we present two frameworks, Ubuntu and maximum feasible participation. Finally, we demonstrate how employing the aforementioned frameworks in AI can benefit society by flattening the current top-down social hierarchies as AI is currently being utilized. Implications are discussed.