An LLM-Based Interview System for Extracting Resilience Factors

Abstract

In socio-technical systems that constitute the foundation of social infrastructure, safety is a critical consideration, and safety measures have been proposed to enhance resilience, which refers to the ability to maintain and recover functionality under changing conditions. Interviews are an effective method for understanding resilience along with its underlying factors. However, they require considerable human and time resources. To address this issue, this study proposes an automated interview method utilizing a large language model (LLM) to efficiently extract resilience and its background factors.