This working paper examines how artificial intelligence (AI) transforms escalation dynamics by shifting the primary transmission mechanism of conflict from material interaction to perception-driven amplification. Building on the Multi-Layer Coupled Complexity Model (MCCM), the analysis conceptualizes escalation as a threshold-based systemic process centered on the loss-of-control threshold (LoCT). The paper develops a mechanism-based framework of “theater effects,” in which low-cost actions are amplified through visibility, narrative construction, and algorithmic distribution, generating disproportionate system-level impacts. It introduces the Information System Amplification Index (ISAI) as an operational proxy for measuring amplification intensity within the information domain. Using a cross-domain, multi-case validation strategy, including the Russia–Ukraine war, the Red Sea crisis, and the Strait of Hormuz, the study demonstrates the empirical presence of perception–impact decoupling and LoCT compression under conditions of high information density. The findings suggest that escalation must be understood not solely as a function of force, but as an interaction between information systems, perception, and institutional capacity. The paper further outlines policy implications for restoring control under AI-driven escalation environments, emphasizing verification, decision latency, and institutional resilience.
Shaoyuan Wu (Fri,) studied this question.