Artificial Intelligence and Digitalisation in the Construction Industry: A Scientometric Analysis of Structural Integration, Thematic Acceleration and the Sustainability Integration Gap

Purpose – This study maps the intellectual structure of research on artificial intelligence (AI) and digitalisation in the construction industry, moving beyond descriptive science mapping to a network-analytic diagnosis of how the field is structurally integrated, how rapidly its themes are accelerating, and whether its sustainability dimension is structurally embedded in the field. Design/methodology/approach – A corpus of 613 Scopus-indexed documents (2016–2026) was analysed through bibliometric performance analysis, keyword co-occurrence mapping in VOSviewer, and a reconstructed co-occurrence network subjected to multi-measure centrality analysis, cluster-cohesion estimation, modularity computation, and temporal emergence detection. Findings – The field has grown at a compound annual rate of 39.6 per cent to a corpus h-index of 73. Betweenness-centrality analysis reveals a hub-and-spoke architecture in which construction industry is the singular structural bridge (betweenness 0.98) despite high network density (0.77). Cluster-cohesion analysis uncovers a descending integration gradient across the field's four thematic clusters, from a well-integrated digital-BIM core (cohesion 0.561) to a barely-integrated sustainability cluster (0.106). Emergence analysis shows that machine-learning and deep-learning methods are accelerating most rapidly, while sustainability themes are both the least cohesive and among the slower-accelerating. The study terms this divergence the sustainability integration gap. Originality/value – This is the first scientometric study of AI and digitalisation in construction to combine multi-measure centrality, cluster-cohesion estimation, and emergence detection, and the first to demonstrate empirically that the field is deepening technologically while leaving its sustainability dimension structurally detached.