ABSTRACT This essay analyzes artificial intelligence not as a homogeneous technology but as an architecture of unequally distributed capabilities that reconfigures operative agency and the exercise of power. The central variable is not the algorithm or model size, but the capacity to close complete cycles of information, decision, and action without institutional friction or human deliberation in relevant time. From that variable, the text constructs an access map describing the effective capabilities of the state, corporations, security forces, citizen collectives, and individuals. It shows how technical inequality becomes power asymmetry when certain actors can define the environment in which others act — without issuing explicit commands and without passing through deliberation. It examines the role of legality not as a substantive limit but as a stabilization infrastructure for the emerging order, and frames the central problem in republican terms: the technical impossibility of controlling, in relevant time, systems that are cognitively superior to the classical mechanisms of citizen scrutiny. The essay identifies the structurally irresolvable contradictions of the system — between speed and control, between symmetry and security, between access and order — and analyzes possible attenuations while making their costs and second-order effects explicit. It proposes no solutions or ideal restorations. The central thesis is that the primary risk is not autonomous artificial intelligence, but the emergence of an order that no longer requires functional citizens.
Javier Ignacio Janer Tittarelli (2026) studied this question.