Abstract
Emerging artificial intelligence (AI) technologies have begun to play a disruptive role in representing interior environments. These AI tools allow users to utilize artificial neural networks to generate and manipulate static spatial ideations. However, as AI-produced images of surreal and extraordinary interior environments become increasingly common in journalistic media and online, designers must question the tool’s applicability in the discipline of interior design. What are the potential effects of this technology, and does the application provide a falsified representation of interior design’s core ideologies and pedagogical foundations? This research seeks to determine if AI software can generate digital images of interior environments that accurately represent interior design’s seven foundational elements: colour, form, light, line, pattern, texture, and space. Structured text prompts based on these elements are input into AI platforms such as Dall-E-2, Interior AI, Stable Diffusion, and Midjourney. The image outputs from each platform are visually analyzed using the aforementioned elements. The investigation argues that current AI image generation tools cannot accurately represent interior space or consistently introduce finer details related to form, texture, line, and lighting. However, the AI tools successfully recreated colour, pattern, and space in some instances. As journalistic media and digital media disciplines rely further on AI to ideate spatial representations, and as AI moves towards producing 3D objects and spaces, we must rigorously question to what extent using a tool lacking these core pedagogical principles means for the representation of interior environments.
Presenters
Jason ShieldsAssistant Professor, Interior Design, University of Manitoba, Manitoba, Canada
Details
Presentation Type
Paper Presentation in a Themed Session
Theme
KEYWORDS
Artificial Intelligence, Interior Environments, Interior Representation, Artificial Intelligence Ethics
