Algorithmic Decisionism: The Transfer of Sovereignty in AI-Based Predictive Intelligence Systems in the Post-2020 Era

Abstract

This study examines the transfer of AI-based predictive intelligence systems from the state to the private sector in the post-2020 period within the framework of sovereignty theory. Its central claim is that this transfer is not merely an operational outsourcing but a delegation of epistemic sovereignty, meaning the authority to define what counts as a “threat” and who counts as an “enemy” has passed to private actors. The claim is constructed through the concept of “algorithmic decisionism,” developed at the intersection of Carl Schmitt’s theory of decisionism and Yanis Varoufakis’s technofeudalism. Using qualitative document analysis and comparative case study methodology, the thesis is tested through four cases. The first is the Lavender, Gospel and Where’s Daddy? systems used by the Israeli army in Gaza. The second is the Maven Smart System used in the U.S. intervention against Iran and the Minab girls’ school strike. The third is Operation Resolve against Venezuela. The fourth is the structural transformation of the new-generation defense software market accelerated by Anduril Lattice. Findings show that target identification, friend-enemy distinction, and preemptive decision-making are structured through the “decision architectures” of private sector actors such as Palantir Technologies and Anduril Industries. The study extends the argument by tracing how world models, the generative simulation layer beneath predictive intelligence, push epistemic sovereignty one level deeper into the ontological grammar of the real. It also examines what algorithmic decisionism implies for middle-power states such as Turkey. The study contributes theoretically to the literature by discussing the crisis areas this structural transformation opens for accountability, democratic oversight, and sovereignty.