URA Thèse – Gizem Yüksek

Cognitive issues and instrumentation of architectural ideation by mixed "text-to-image" - "image-to-image" generative AIs

Artificial intelligence (AI), both text-to-image and image-to-image, is currently the subject of much enthusiasm across all disciplines. This mixed generative AI technology enables the creation of images from text or images, thus arousing keen interest in the ideative phases of architectural design, among others.

Our doctoral research explores these contemporary technological developments. More specifically, it aims to gain a detailed understanding of the issues involved in using multimodal AI to instrument the ideative phases of architectural design.

Although these technological advances may seem promising, there is still a lack of research on how these tools influence architects' cognitive processes, particularly during the ideation phase.

In particular, this research examines the dialogue between internal representations, encompassing individual cognitive structures, and external representations, manifested through physical media containing information. In parallel, our aim is to develop and experiment with a mixed generative AI tool aimed at instrumenting architectural design.