Automating HFACS-Based Mishap Analysis Using AI-Assisted Text Classification: 23rd International Conference, EPCE 2026, Held as Part of the 28th HCI International Conference, HCII 2026, Montreal, QC, Canada, July 26–31, 2026, Proceedings, Part I

Accident and incident investigations rely heavily on structured human factors frameworks such as the Human Factors Analysis and Classification System (HFACS). While HFACS provides a robust taxonomy for analysing unsafe acts, supervisory factors, and organisational influences, its application remains predominantly manual, time-intensive, and analyst-dependent. This paper presents an AI-assisted, web-based approach for automating the extraction and preliminary classification of mishap report narratives into HFACS categories. The proposed system integrates document processing, rule-based natural language pattern matching, and structured data storage to support rapid HFACS pre-classification from official accident investigation reports. Using PDF source documents, textual content is extracted and analysed against HFACS-aligned semantic indicators, producing a structured representation of potential human and organisational factors. A user-facing interface enables controlled upload, processing, and review of classified outputs. Rather than positioning automation as a replacement for expert judgement, the system is designed as a decision-support and triage tool, accelerating early-stage analysis and reducing cognitive workload for investigators. Evaluation using representative aviation accident reports demonstrates the feasibility of automating HFACS-oriented structuring while highlighting critical limitations related to semantic depth and contextual interpretation. The findings support a hybrid AI-human workflow for future accident investigation support systems.