Abstract

The fusion of digital advancements and biological systems is transforming modern-day architecture, and this bio-digital approach, when paired with AI image generation models, promises novel design possibilities. The major drawback of this merger is the dismal performance of AI text-to-image models in translating organic tectonic details into architecture. This study examines the complexity of processes, materials, and techniques necessary in a bio-digital architectural approach. Through a series of digital trials, it identifies the need for sophisticated computational models that can capture the complex intricacies and subtle nuances present in living organisms. Before the testing, a set of parameters considered the limitation of how much tectonic information an image could portray. The Nautilus Shell, ferns, mushrooms, seahorses, and grasshoppers were taken as inspiration models because of varying biological configurations. Next, two AI image-generating tools, Midjourney and Stable Diffusion, were used with three different prompt types, each with varying degrees of complexity drawn from five organic systems. A critical analysis of AI-generated images led to the conclusion that, despite AI's exceptional abilities in creating visual content, the complex comprehension of biological systems and their conversion into architectural designs faced significant challenges.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call