AIC-BASED UNFORESEEN EMERGENT SCENARIOS (BLACK SWAN EVENTS) DISCOVERY METHOD FOR AUTONOMOUS SYSTEMS SAFETY DESIGN

Abstract

This paper presents a new approach to enhancing the safety design of autonomous systems by applying systems science in identifying autonomous systems-related Unforeseen Emergent Scenarios (Black Swan events)—unexpected situations requiring critical decisions. Building on the AIC (Appreciation, Influence, Control) systems approach, this research incorporates a perspective shift to better analyse potential system failures in complex environments. The study introduces a systems thinking method with an AIC-based Chain-of-Thought (CoT) technique, which helps architects predict and analyse autonomous unforeseen behaviours that could lead to failures. By altering the type and direction of AIC interactions, architects can improve their predictive capabilities and prepare for complex interactions that traditional methods may overlook. The result is a list of Unforeseen Emergent Scenarios that can inform additional safety requirements, laying a strong foundation for designing safer autonomous systems.