Artificial Intelligence in Architectural Design: Scientometric Mapping and a Futures Typology

Artificial intelligence (AI) is increasingly embedded in architectural design through generation, prediction, simulation, optimization, and data-driven decision support. Yet research on AI in architectural design (AI-in-AD) remains dispersed across architectural computation, building-performance optimization, construction automation, energy research, and wider AI domains. This study provides an AD-delimited scientometric analysis of 2541 Scopus-indexed publications spanning 1974–2026, retrieved on 5 September 2026; coverage of 2026 is partial. Publication trends, keyword co-occurrence, document co-citation, citation-burst detection, direct citation among publication sources, and institutional and country/region co-authorship networks were used to characterize the field’s historical, thematic, intellectual, publication-source, and collaborative structures. The results show rapid expansion after 2021 and identify three interconnected thematic domains: AI-enabled generative and representational design workflows, performance-driven optimization, and machine-learning-based sustainability and intelligent-building research. Co-citation and citation-burst patterns indicate a succession from early space-planning, expert-system, and evolutionary approaches through building-performance optimization to recent deep learning and generative AI applications. Publication and collaboration structures span multiple disciplinary outlets, while research output and collaboration patterns are unevenly distributed across institutions and countries/regions. Building on these findings, this study develops a conceptual futures typology comprising sixteen configurations defined by four identified conditions: direction, pace, purpose, and distribution. The typology distinguishes alternative futures relative to the field’s current trajectory.