AI agents that generate final answers based on user input often fall short in creative fields like structural design and architecture. These domains require interactive systems that assist users in externalizing and developing ideas, exploring alternatives, and refining partial solutions.
The final product must adhere to various constraints concerning spatial configuration, mechanical behavior, material quantities, and costs. These constraints create friction in the design process, which can stimulate novel and creative solutions.
This paper discusses the misalignment between current generative AI goals aimed at removing friction and the needs of creators, such as structural designers, who develop ideas through iterative work. We present design dimensions for systems that facilitate constrained human-AI co-creation, leveraging vision-language models to make structural exploration conversational, multimodal, and responsive to evolving human intent, thus enhancing the creative process.
Through a pilot design interface based on these principles and a study with industry experts, this paper illustrates how structural designers perceive interactive AI systems and how these systems can support design space exploration by reducing repetitive modeling friction while preserving reflective design friction.
Blogger's Review: This paper delves into the significance of human-AI collaboration in structural design, highlighting the constructive role of friction in the creative process. By introducing vision-language models, the study reveals how to achieve interactivity in design, offering fresh perspectives that are worth exploring and learning from.