The Necessity of Sovereign AI and the Governance of National Security Risks

A Framework Drawn from Lorenz von Stein's Theory of the State

Abstract

As states increasingly depend on foreign technology providers for algorithms, computational infrastructure, and training data, “sovereign AI” has emerged as a national security priority alongside export controls and data governance. Existing scholarship, however, largely reasons at the level of industrial policy and geopolitics, leaving the state-theoretical foundations of AI infrastructure governance under-examined. This article develops a doctrinal and comparative analysis grounded in Lorenz von Stein's nineteenth-century theory of the administrative and social state, applying his account of state personality and self-preservation to computational infrastructure, and comparing Taiwan's Cyber Security Management Act and AI Action Plan with the European Union's AI Act and India's Digital Public Infrastructure initiative. Sovereign AI, reconstructed through Stein's framework, is best understood as a contemporary exercise of the administrative state's obligation to maintain social integration and self-preservation. Legal design should embed computational infrastructure, model autonomy, and data sovereignty into national security law, while guarding against techno-nationalism and surveillance concentration.