The Large Design Model (LDM): An Integrative Review of the Convergence of Generative AI and Architectural Design (2015–2026)
The Large Design Model (LDM): An Integrative Review of the Convergence of Generative AI and Architectural Design (2015–2026)
Saman Mohri1
1) Adjunct Lecturer, Department of Architecture and Urban Planning, Technical and Vocational University (TVU), Shiraz, Iran,
IOI Code
:
AACZ-ABCZEC
Publication :
The 5th International Conference on Seismic Retrofitting(5icst.ir)
Abstract :
The accelerating integration of generative artificial intelligence into architectural design is fundamentally transforming the profession. Despite remarkable advances in large language models (LLMs), generative vision models, and agent-based systems, the absence of a unified framework for understanding the convergence of these three technological pillars represents a significant theoretical gap. Using Whittemore and Knafl s integrative review methodology and critical analysis of 50 selected sources from Scopus, Web of Science, and Google Scholar (2015–2026), this study proposes the conceptual model of the Large Design Model (LDM). Built upon the analogy of Linguistic Brain, Creative Eye, Executive Hand, the LDM features a four-layer system architecture comprising multimodal input, convergent processing, agent-based decision-making, and multi-scale output. Critical analysis reveals that each technological pillar alone faces fundamental limitations, and only three-way convergence can address architectural design complexity. Three categories of technical, epistemological, and ethical challenges are identified, and five strategic research priorities for 2026–2030 are proposed.
Keywords : large design model generative ai large language model diffusion model agent based system architectural design