Generative Artificial Intelligence (GenAI) text-to-image (TTI) technology, like Midjourney, can create visually impressive images from text prompts but does not fully capture the creative synthesis essential to design and architecture. This study examines the concept-synthesizing abilities of fifteen professional architects and designers using GenAI TTI. Verbal protocol analysis was conducted in a three-stage experiment: (1) concept synthesis with conventional tools, (2) refinement using GenAI TTI, and (3) interviews to validate the findings. Data were analyzed using word association to assess tendencies toward convergent or divergent thinking, with “Indonesian Identity” as a key parameter. Results revealed three creative modes: (1) Co-visualize mode, where GenAI TTI acted as a visualization tool; (2) Co-wander mode, where GenAI TTI was an experimental medium encouraging exploration; and (3) Co-synthesis mode, where GenAI TTI served as a creative partner, fostering complex ideation in ambiguous spaces. The co-synthesis mode facilitated high-level conceptualization, allowing access to dimensions beyond conventional approaches.
Pebriani et al. (Wed,) studied this question.