Application of large language models in construction methods management

Abstract

In the context of digital transformation, the preparation, inspection, and management of construction methods still mainly depend on individual experience, which can lead to errors and inconsistencies. Therefore, this research aims to explore the potential of applying Large Language Models (LLMs) to enhance the efficiency of construction method management in the construction industry. The study examines core functions of LLMs, such as natural language processing, information synthesis, decision support, and automation of technical documents. Based on this, the paper proposes a model for applying LLMs in key stages, including: (1) supporting the preparation of construction methods; (2) reviewing technical and safety content; (3) managing knowledge and standards; and (4) assisting with training and capacity building for engineers.
The research findings show that applying LLMs reduces time, improves the accuracy of construction methods, and effectively supports risk and occupational safety management. However, the paper also highlights challenges such as data reliability, the need for expert verification, and information security issues. Therefore, the study proposes a strategy for deploying LLMs in construction management that integrates technology and human expertise to drive digital transformation and enhance the quality of construction project management in Vietnam.