Algorithmic Sovereignty: A Multidimensional Conceptual Framework for National Cyber-Defense Posture in Emerging Economies

Abstract

The accelerating embedding of artificial intelligence (AI) into national cyber defense is quietly reconfiguring what it means to be sovereign in cyberspace. Prevailing conceptions of cyber and data sovereignty privilege control over data and physical infrastructure, yet they largely bypass the algorithmic layer—the foundation models, inference pipelines, and machine-learning systems on which contemporary defensive operations increasingly rely. This paper develops algorithmic sovereignty as a multidimensional construct denoting a state’s capacity to exercise autonomous and meaningful control over the AI systems embedded across its cyber-defense lifecycle. Integrating the resource-based view, dynamic-capabilities theory, securitization theory, dependency theory, and the literature on cyber power, we theorize six interdependent dimensions—infrastructural, data, model/algorithmic, operational, capability, and governance/normative sovereignty—and cross them with a five-stage maturity continuum to produce a diagnostic matrix. We advance five testable propositions, including the claim that data localization without model-layer control yields a sovereignty illusion, and that absorptive capacity moderates the dependency–vulnerability relationship. Because the contribution is conceptual, we specify a four-stage validation protocol—systematic review, Delphi-based content validation, multi-criteria weighting, and illustrative national instantiation—and operationalize each dimension into a candidate indicator bank suitable for empirical testing. The framework is bounded to middle-power and emerging-economy contexts, where an autonomy–capability trade-off is most acute. We discuss theoretical and policy implications using Indonesia as an illustrative case and outline an agenda for empirical validation.