The Manhattan Trap: Why a Race to Artificial Superintelligence is Self-Defeating

TL;DR

The paper analyzes the strategic incentives behind a potential race to Artificial Superintelligence (ASI) and argues that the same assumptions pushing for a race make it extraordinarily dangerous. Using a defensive realism lens, it identifies three core risks—great power conflict, loss of control, and domestic power concentration—that undermine stability and liberal democracy. It reframes the scenario as a trust dilemma rather than a prisoners’ dilemma, arguing that credible verification and governance can enable cooperative restraint. The authors conclude that international cooperation, supported by feasible verification regimes and political will, is preferable and strategically sound to avoid the existential risks of racing to ASI, with significant implications for security policy and AI governance.

Abstract

This paper examines the strategic dynamics of international competition to develop Artificial Superintelligence (ASI). We argue that the same assumptions that might motivate the US to race to develop ASI also imply that such a race is extremely dangerous. These assumptions--that ASI would provide a decisive military advantage and that states are rational actors prioritizing survival--imply that a race would heighten three critical risks: great power conflict, loss of control of ASI systems, and the undermining of liberal democracy. Our analysis shows that ASI presents a trust dilemma rather than a prisoners dilemma, suggesting that international cooperation to control ASI development is both preferable and strategically sound. We conclude that cooperation is achievable.