Description
Abstract:
The accelerating diffusion of artificial intelligence (AI) capabilities across various domains of strategic significance such as military, economic and information, has introduced a structural tension at the core of great power politics. The states, here, simultaneously compete to establish AI dominance while recognizing the systemic risks that unregulated AI development poses to collective stability in the international arena. This paper examines the condition wherein the same AI capabilities that incentivize arms race and great power rivalry also generate the shared vulnerabilities that make cooperation among the states, both necessary and politically viable.
Drawing on Keohane and Nye's complex interdependence framework alongside Baldwin's relational conception of power, this paper argues that AI does not merely replicate the Cold War deterrence logic. Instead, it produces a novel strategic environment characterized by challenges of opacity, speed and domain-diffuseness that raise the costs of conflict while simultaneously complicating the nature of confidence building measures (CBMs) essential for cooperative regimes. The paper applies this framework comparatively across the United States, China, the European Union and other major AI powers, demonstrating that AI is simultaneously concentrating frontier capabilities among a narrow cluster of actors while enabling selective asymmetric leverage for the some actors pursuing strategic autonomy. The paper then concludes by assessing whether meaningful AI governance agreements are achievable among the US, China, and EU and under what conditions the alignment choices of other actors become structurally determinative in shaping the power trajectory of these actors.
Keywords: Artificial intelligence, Arms Race, Great Power Rivalry, Complex Interdependence, Domain-Diffuseness, Relational Power, Strategic Autonomy