Abstract
This paper reflects on the impact of recent developments in generative AI and the increased accessibility of text-to-image models on the role and status of the image in architectural practice and culture. Since its emergence as a profession and discipline architecture has primarily on visual representation, in particular architectural drawing, as a means of envisioning, designing and constructing built environments. In contrast to other artistic and designerly disciplines architects do generally not work on the object they design directly, but through some intervening medium, i.e. architects don’t make buildings but representations. Architectural drawings are orthographic: they represent, existing or imagined, spatial constructs into two dimensional notation through orthographic projection. The last decades with the emergence of technologies such as computer aided design, building information modelling, we have seen shift from an analogue, two dimensional, drawing based approach towards a digital mode, three dimensional modelling based approach, in architectural design. While architectural culture has not come to terms yet with this shift from representation to simulation, from orthographic drawings to post-orthographic images, generative AI seems to further push the synthetic nature of images. Generative AI tools trained on vast amounts of data rapidly accelerate the shifts outlined above, raising questions on how we include AI in architectural design practice but also on the value of images, and how we attribute authorship. This paper builds on a series of experiments in synthesizing architectural images in the context of architectural education.
Details
Presentation Type
Paper Presentation in a Themed Session
Theme
2024 Special Focus—Images and Imaginaries from Artificial Intelligence
KEYWORDS
Architecture, Imaging, Representation, Simulation, Synthesis, Generative ai
