Perceived and actual AI literacy in military organizations: a self-efficacy framework for training design

This study examines the underlying causes of the gap between perceived artificial intelligence (AI) literacy and the actual competencies of defense personnel, and explores structured training initiatives that can support reliable and responsible decision-making in AI-mediated operational environments. Grounded in self-efficacy theory and the cognitive psychology of competence assessment, the study advances the theoretically motivated proposition that the routine use of intelligent systems may contribute to overconfidence and illusions of competence, dynamics that, while documented in educational and civilian professional contexts through existing research on automation bias and self-efficacy inflation, have not yet been systematically examined in military organizational settings and whose relevance to high-stakes defense environments constitutes a central theoretical proposition of this study. Drawing on a narrative review of the literature, the study analyzes the cognitive, metacognitive, and organizational factors contributing to this gap, and surveys AI literacy initiatives adopted by major defense organizations, highlighting a tentative cross-national convergence toward a multidimensional model integrating cognitive, operational, critical, and ethical competencies. Building on these insights, the authors present a foundational fourdimensional training framework implemented at the Center for High Defense Studies/School of Advanced Defense Studies (CASD/SSU) as an illustrative programmatic case, demonstrating how structured AI literacy pathways may be designed to promote awareness, calibrated self-efficacy, and organizational readiness in technologically complex military environments.