Representing the spatial projection of mental imagery through generative artificial intelligence: The Atlas of Misty Continents case

Abstract

Design is the transformation of an image in the mind into a final product. The client and the designer each have an image of this product. Clients often use words to describe this image. The designer understands the wishes, demands, and needs expressed in words and brings these images as close as possible. Through aesthetics, functionality, ergonomics, and feasibility filters, s/he transforms the image into physical form. In this process, many tools can be used. Generative AI is one. The study aims to explore the possible contributions of image-to-image transformation to the visualization of space and develop a flowchart in which different generative AI approaches participate. For spatial review and schema production, the area of literature where words represent space was used. A space model that is depicted in İhsan Oktay Anar’s The Atlas of Misty Continents has been used. A mixed methodology is used in the study. Databases and search engines were scanned using quantitative methods, and the scope of the study was drawn by determining the models to be studied through criterion sampling. Content analysis was used to examine the outputs and models. Geometric inconsistencies were found between the source images and the final outputs. As a result, it was seen that although many possibilities were offered to designers in the model interfaces, the models were not yet sufficient for space visualization. The study will contribute to ongoing discussions about the position and benefits of generative AI in design.