Artificial intelligence is increasingly embedded in military decision-making systems, transforming not only operational performance but the structural conditions under which risk is produced and governed. This article develops a conceptual socio-technical analysis of AI-enabled automation through the lenses of systemic risk, risk governance, and resilience. Rather than treating AI as a discrete tool, the paper conceptualizes automation as a structural intervention in high-risk decision infrastructures. Four analytical findings are advanced. First, AI-mediated workflows redistribute practical control by shaping what information is salient, how threats are classified, and which actions appear feasible. Second, automation compresses decision time, reducing space for deliberation and increasing reliance on algorithmic outputs under uncertainty. Third, tighter coupling and accelerated interaction across domains enable localized errors or manipulations to propagate into broader escalation dynamics. Fourth, distributed human–machine architectures contribute to responsibility diffusion, weakening the alignment between accountability and effective decision influence. The article argues that prevailing “human-in-the-loop” safeguards are insufficient when control is reduced to technical override rather than institutional capacity for judgment and accountability. Effective integration of AI in high-risk environments requires governance arrangements that preserve human judgment as a strategic resource for crisis management and systemic resilience.
Authors
Dejan Vuletić
Strategic Research Institute, University of Defence, Veljka Lukića Kurjaka 33, 11042 Belgrade, Serbia
Author
https://orcid.org/0000-0001-9496-2259
DOI: https://doi.org/10.66050/xhmjy003
Keywords: artificial intelligence, military decision-making, risk governance, socio-technical systems, escalation risk
https://internationaljournalofdisasterriskmanagement.com/index.php/Vol1/article/view/198
