Abstract
Due to the rapid advancement of technology and artificial intelligence (AI), there has been an increase in the number of studies looking into pedagogical innovations relating to emerging technologies in this area. Text-to-image AI-driven approaches supported by diffusion models and Natural Language Processes (NPLs) like Midjourney, DALL E 2 and Playground are poised to alter the technological landscape for architecture students, becoming an important component within a student design process and affecting creative decision-making methodologies. However, architectural education has yet to fully comprehend the full extent of these new models, their potentials and limitations for creative and professional practice, as well as challenges and pitfalls for the next generation of architects and designers.
In our Master of Architecture programme, we have been experimenting with diffusion models and new ways of generating images with very promising results. In this paper, we present a theoretical framework where we outline the challenges in developing an AI-driven student project. We present the findings of an AI workshop where each group of students used Midjourney, DALL E 2, and Playground to generate architectural scenarios using textual prompts. The results are used to speculate on the agency of designers while using AI-assisted processes, suggesting new possible pathways for future pedagogy in architecture and design.
More generally, this study discusses the use of AI in architectural education and suggests areas requiring more research and potential future applications. It also explores some pedagogical implications of AI use within architectural education that strike a balance between creativity, deterministic approaches, algorithmic logics and human skills. The findings of this study are expected to address the conference themes of boundaries of expertise; collaboration, co-production, power and agency; multi and inter-disciplinary pedagogy and practice.
In our Master of Architecture programme, we have been experimenting with diffusion models and new ways of generating images with very promising results. In this paper, we present a theoretical framework where we outline the challenges in developing an AI-driven student project. We present the findings of an AI workshop where each group of students used Midjourney, DALL E 2, and Playground to generate architectural scenarios using textual prompts. The results are used to speculate on the agency of designers while using AI-assisted processes, suggesting new possible pathways for future pedagogy in architecture and design.
More generally, this study discusses the use of AI in architectural education and suggests areas requiring more research and potential future applications. It also explores some pedagogical implications of AI use within architectural education that strike a balance between creativity, deterministic approaches, algorithmic logics and human skills. The findings of this study are expected to address the conference themes of boundaries of expertise; collaboration, co-production, power and agency; multi and inter-disciplinary pedagogy and practice.
Original language | English |
---|---|
Title of host publication | Association of Architectural Educators Conference 2023 |
Publication status | Submitted - 15 Nov 2022 |