This chapter develops a philosophical and technical framework for AI Pluralism—the design of AI systems as networks of diverse intelligences capable of sustaining multiple perspectives and negotiating toward provisional solutions. Drawing on John Dewey’sDewey democratic experimentalism, Donna Haraway’sHaraway situated knowledgessituated knowledges, and G.W.F. Hegel’s dialectical method, we argue that when confronting morally complex and epistemically contested questions, truth emerges through structured interaction rather than monolithic optimization. We introduce FOAM (Framework for Openly Augmented Mediation), a computational architecture that operationalizes these philosophical insights through three core primitives: diverse agents with explicit stance vectors, deliberative protocolsdeliberative protocols for structured critique, and sublation operators that weave contradictory perspectives into emergent syntheses. Using the fictional Port AquiloPort Aquilo climate council as a recurring example, we demonstrate how pluralistic AI can preserve productive disagreement while generating actionable outcomes. The chapter addresses key tensions in pluralistic design—richness versus clarity, memory versus revision, transparencytransparency versus privacyprivacy, and computational cost versus civic worth—arguing that these trade-offs are features, not bugs, of democratically legitimate AI systems.