Unpacking trends in artificial intelligence research in the construction industry: a bibliographic review

Résumé fourni par la source

Introduction The study examines the evolution, intellectual structure and global distribution of artificial intelligence (AI) research within the construction industry, with the aim of systematically mapping scholarly trends and identifying future research directions in a rapidly maturing field. Methods A Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guided bibliographic review was conducted using the Scopus database. An initial dataset of 1,489 publications was systematically screened, resulting in 764 journal articles and conference papers published between 2015 and 2025. Bibliometric techniques, including productivity analysis, citation analysis, and keyword co-occurrence network mapping using VOSviewer, were applied to examine research growth patterns, influential contributors, and dominant thematic clusters. Results The results reveal sustained and accelerated growth in AI-related construction research, particularly after 2020, indicating increasing conceptual consolidation. High-impact journals and a small group of leading authors and institutions dominate knowledge production. Four major thematic clusters emerge: AI-enabled safety and automation, predictive modelling and optimisation, digital life-cycle integration, and AI-based decision support. While the United States, United Kingdom and China lead global output and citation impact, several developing countries also demonstrate substantial research participation. However, citation influence and global visibility remain concentrated within established research systems. Discussion The findings demonstrate that AI research aligns closely with core industry challenges, particularly safety management, performance prediction, and data-driven decision-making. Construction organisations are likely to derive the greatest value from AI when it is embedded within integrated digital information environments such as BIM and digital twins. The study is limited to English-language, Scopus-indexed publications and does not assess real-world implementation outcomes. Future studies should incorporate multiple databases and empirical case analyses, particularly in underrepresented regions. The study provides one of the most comprehensive, PRISMA-compliant bibliographic syntheses of AI research in construction, offering a structured intellectual map of a rapidly maturing field.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.