An AI-supported Method for Scalable Micro-Exercises for Crisis Training

Abstract

This article presents an ongoing study on AI-supported micro-exercises for crisis management training. The research is conducted across three municipalities and incorporates insights from educational contexts. It involves parallel tabletop exercises simulating Emergency Operations Centres (EOC). Audio recordings from these exercises were transcribed using OpenAI Whisper and analyzed with AI tools such as ChatGPT. The AI-generated transcripts facilitated comparisons of the groups' working methods and enabled reflective organizational learning.

Preliminary findings indicate the potential of AI to offer a scalable, data-driven exercise format that supports documentation, analysis, and after-action review. Our study advocates the advantages of flexible and repeatable micro-exercises and underscores the importance of ethical considerations when employing AI in crisis training.