Advancing Construction Innovation: Bibliometric Insights into Large Language Models in the Construction Industry

Large Language Models (LLMs) have revolutionized industries worldwide, and the construction industry is no exception. LLMs enhance digital solutions for construction design and management. It further promotes stakeholder collaborations and assists in decision-making by processing large datasets and evaluating embedded systems in modular designs. This study explores the impact of LLMs in the construction industry through a bibliometric analysis of 24 documents retrieved using the Elsevier Scopus database with keywords “large,” AND “language,” AND “models,” AND “construction,” AND “industry” spanning 2000 to 2024. Using a VOS viewer, the research maps the bibliometric relationships among these documents to uncover key themes, trends, and research gaps in applying LLMs in construction. The analysis identifies four clusters with emerging themes, including Digital solutions for Construction Design and Management, Systems Engineering and Modular solutions for Sustainable Development, AI-driven Language Processing in Construction modelling and Automated Information Processing and Compliance in Large Datasets. The findings also reveal significant gaps in research. Despite the evident potential of LLMs in streamlining construction industry processes, there is a substantial research gap in addressing the customization and domain-specific adaptation of LLMs to meet the specific requirements of construction industry tasks. Existing studies primarily focus on generic applications of LLMs, such as information retrieval and data processing, but lack exploration into their tailored integration for complex tasks like regulatory compliance, modular construction optimization, and sustainable development. Furthermore, geographic limitations with the United States of America and China leading in research in existing literature highlight a lack of studies focused on developing countries, where the industry is rapidly growing but struggles with adopting digital innovations like LLMs. While the study provides valuable insights, it is limited by the relatively small dataset of 24 documents and the narrow focus of the Scopus search criteria. Future research could expand the dataset by including broader keywords or alternative databases and examine deeper into cross-regional comparisons. Notwithstanding these limitations, the study significantly contributes to the growing body of knowledge in understanding the integration of LLMs in the construction industry and provides a foundation for further exploration.