Exploring text-to-image application in architectural design: insights and implications

This study explores the potential of employing AI, particularly text-to-image generators, in the architectural design process. It addresses three main research questions: (1) How do designers utilize text-to-image generative models in architectural design practices? (2) From a designer's perspective, what are the most significant limitations and potential threats associated with using text-to-image generative models? (3) What is the role of text-to-image generative models in enriching the design process? Using a qualitative research method, semi-structured interviews were conducted with sixteen experienced architectural designers who had recently incorporated text-to-image generative models into their practice and toolkits. The findings reveal that these models have the potential to enhance creativity, visualization, and imagination, particularly in the early design stages. However, participants also identified difficulties, limitations, and potential threats, emphasizing the need to improve the tools further to fit the architectural design field. The results of this study are demonstrated as a matrix that illustrates the different tools used by architects and designers during the design process and the addition of text-to-image models to these tools. Moreover, the research findings are summarized using a SWOT analysis, outlining the strengths, weaknesses, opportunities, and threats associated with incorporating text-to-image generative models in architectural design. This study will help designers in the architectural design field, including professionals, academics, and students, to boost their creative process when designing projects and empower them to be more productive in their specializations by facilitating the augmentation, exploration, and experimentation of new ideas.