An Integrated Theoretical Framework for Integrating Generative Artificial Intelligence in Architectural Design Studios: From Programming to Schematic Design

An Integrated Theoretical Framework for Integrating Generative Artificial Intelligence in Architectural Design Studios: From Programming to Schematic Design

Kamran Javaheri1 Hana Avani2

1) Department of architecture, Se.C, Islamic Azad University, Semnan, Iran,
2) Department of architecture, Se.C, Islamic Azad University, Semnan, Iran,

IOI Code : AAHF-ABFGAF
Publication : National Conference on Architecture and Art Technologies, Smart City, Soft and Creative Industries and Technological Culture(icact.ir)
Abstract :
    The present study aims to develop and validate an integrated theoretical framework for utilizing generative artificial intelligence across the entire architectural design process, from programming to schematic design. Despite the rapid proliferation of generative AI tools in recent years, the existing literature lacks a framework capable of bridging the theoretical gap between machine output and architectural design logic. Employing an operational design research approach and examining two real architectural projects of varying complexity, this study introduces and tests the four-phase AIPD framework. The findings demonstrate that the proposed framework can address all three major theoretical gaps: the translation gap (from program to spatial organization), the feedback gap (cross-scalar oscillation), and the judgement gap (qualitative evaluation). For the low-complexity project, the framework performed at a level equal to or better than conventional manual processes, while for the high-complexity project, it achieved the same qualitative level in a significantly shorter time. The most significant theoretical contribution of this research is demonstrating the feasibility of human-machine collaborative design in architecture, contingent upon three conditions: the presence of an intermediate adjacency matrix layer, cross-scalar feedback simulation through a feedback agent, and complete delegation of final judgement to the human architect. These findings have direct implications for architectural firms, software developers, and architectural education curricula. This research shows that generative AI will not eliminate architecture; rather, it will transform it so profoundly that the architect s role shifts from form-maker to selective judge of possibilities.

Keywords : generative artificial intelligence architectural design architectural programming schematic design human machine interaction integrated theoretical framework