Using Automatic Speech Recognition for Documenting Work in Municipal Emergency Operations Centers

Abstract

Automatic speech recognition (ASR) and automatic documentation have not been widely explored in crisis management, despite their potential utility in facilitating the transcription of speech recordings. Although documentation is widely recognized as essential for creating a common operational picture, there is often a lack of such documentation, which can hinder understanding of events during and after a crisis. The novelty of the research is to apply existing technology and evaluate the potential of ASR technology in the domain of crisis management. We present preliminary results of using OpenAI Whisper for automatic transcription and documentation. In Phase 1, the ASR software was tested with existing recordings from previous research. In Phase 2, data was collected from recorded meetings during a tabletop exercise. The results indicate that transcripts, combined with other AI technologies, can provide valuable information and support for crisis and emergency management in Emergency Operation Centers.