AI, Conditioned Uncertainty, and the Functions of Elections in Hybrid Regimes

This paper develops a theoretical framework for understanding how artificial intelligence (AI) reshapes the functions and political significance of elections in hybrid regimes. While elections in these regimes remain formally competitive, their outcomes are constrained due to uneven playing field (Levitsky and Way, 2001). Existing scholarship on electoral authoritarianism has illuminated these asymmetries, yet it has not fully accounted for how new digital technologies transform them. Building on insights from research on democratic backsliding and algorithmic politics, I argue that AI functions as an emerging technology of rule that modifies how conditioned uncertainty is produced and managed. AI-driven tools such as bots, recommender systems, automated amplification mechanisms, and synthetic media, do not eliminate electoral uncertainty but recalibrate it by making manipulation more scalable, more opaque, and more adjustable to shifting political conditions. Based on an analysis of several illustrative cases, including Türkiye, India, and the United States, the paper demonstrates how AI-enabled persuasion, disinformation, and narrative shaping are strategically employed to reinforce incumbent advantage, fragment opposition coalitions, and distort public perception of electoral competition. The framework makes three contributions: (1) it conceptualizes AI as a new layer within existing repertoires of manipulation; (2) it theorizes variation in how AI reshapes electoral competition in democratic and hybrid regimes; and (3) it identifies the conditions under which opposition actors can still leverage elections to challenge incumbents despite unbalanced technological advantage. By integrating insights from comparative politics, political communication, and AI governance, the paper advances a new research agenda for understanding elections under illiberal and technologically evolving political contexts.