ASSESSMENT OF THE CURRENT STATE OF AI-AIDED DESIGN IN ARCHITECTURE CASE STUDY: DESIGN OF A RESIDENTIAL BUILDING IN RIYADH CITY

Document Type : Original Article

Authors

1 Architecture and construction sciences, College of Architecture and Planning, King Saud University, Riyadh, KSA

2 Architecture and Construction Sciences, College of Architecture and Planning, King Saud University, Riyadh, KSA

Abstract

In the AEC field, AI-aided design tools have shown to be beneficial during design process from its very beginning stages. It usually requires human intervention as a normal progress for a recently-developed technology in which is referenced as ‘augmented intelligence’.
In this research, the capability of AI-powered tools has been tested regarding architectural design. A residential building in Riyadh City was chosen for a redesign case study. The AI proposal design has been divided into two phases presented in the floor plan layouts and the building façade design. Depending on the type of task to be resolved, a comparison among related AI tools has been implemented based on specific criteria to select the best fit for each phase. GenAI tool produced the layouts, then slightly modified manually to rectify inaccuracies. The result has been evaluated in reference to the requirements parameters and to the original proposal. For façade design, diffusion-based model has been utilized via text and image-to-image translation. The result has been evaluated regarding the accuracy and creativity.
The experiment reveals the current potential of two different types of AI technologies. It is confirmed by the evaluation that AI is still not able to dominate in the field in the current time. However, as seen in ArchiteChtures and MidJourney, this capability varies based on the AI model and the task performed. While it was challenging to reach the optimal spatial layouts, the result of the façade design showcases the high capability of this kind of AI models.

Keywords

Main Subjects