Knowledge-Based Systems in the Era of Large Language Models: A Case Study in Fire Safety Management

Abstract

The construction industry faces complex decision-making challenges that require the integration of extensive domain knowledge and data analysis. Traditional methods are inadequate to effectively address these challenges due to the dynamic nature of construction projects. In the fast-evolving landscape of construction technology, artificial intelligence (AI) plays a crucial role in enhancing decision-making and prediction processes. Therefore, the main purpose of this research is to examine how AI, particularly Large Language Models (LLMs), can be integrated in the analysis and management of emergency scenarios to improve decision-making, prediction, and optimization through knowledge-based systems. This research follows the CommonKADS method for knowledge engineering and embeds sub-symbolic and symbolic AI techniques. By integrating LLMs, this research conceptualizes and operationalizes tacit construction knowledge into structured knowledge systems. This approach has been applied in a case study analysing National Fire Protection Association (NFPA) fire incidents to test possible interactions to extract related data for the knowledge-based systems. The findings inspire future research on the scalability of AI-integrated systems across different segments of the construction industry, their potential in regulatory compliance scenarios, and the development of automated knowledge-based systems in construction industry.