The magic of AI: Unveiling the sociopolitical implications of belief in artificial intelligence

Abstract

This introduction argues that anthropology requires a new conceptual framework for understanding the contemporary social life of artificial intelligence (AI). We propose "the magic of AI" as an analytical concept that shifts attention from AI's technical capacities to the beliefs, uncertainties, and sociotechnical imaginaries through which it acquires authority and efficacy. Revisiting anthropological theories of magic, particularly the work of Ernesto De Martino, Bruce Kapferer, and Francesco Della Costa, we argue that AI's apparent autonomy functions less as a property of the technology itself than as a collectively sustained form of belief. These beliefs naturalize AI's purported inevitability, mobilize political and economic investments, and obscure the human labour, infrastructures, and relations of power through which AI is continually produced. Rather than employing magic as a metaphor, we develop it as a theoretical lens for analysing how opacity, trust, and anticipation generate material sociopolitical effects. In doing so, this introduction offers a conceptual framework for understanding AI as a contingent sociotechnical formation whose power rests as much in collective imagination as in computation itself, while introducing the contributions that develop this argument ethnographically across diverse settings.