2025 ASEE Annual Conference & Exposition

The addition of generative artificial intelligence (AI) algorithms to the traditional engineering design process, e.g., generative design (GD) and large language model (LLM) assisted design, significantly shifts the role, cognition, and behavior of the human designer. The way the human explores the design space is the design task most affected by the generative AI-based paradigm shift. Traditional design (TD) characteristically requires the human designer to explore the design space by generating a range of ideas/potential solutions which might solve the goal(s), and then analyze their performance in the objective space. However, the GD process prompts designers to employ inverse thinking by computationally defining and scoping the objective space (e.g., materials and manufacturing constraints) for the AI-agent to explore and analyze all available options in the design space while optimizing towards (a) feasible solution(s) which are presented to the human designer for further selection and iteration. The inverse relationship of TD and GD requires the designer to think and behave in a unique manner for each paradigm. However, there is little research exploring Generative Design Thinking, i.e., the cognitive processes that underlie GD. There also lacks an available infrastructure for conducting and applying the results of educational research to determine best GD teaching practices and develop new teaching materials.

The goals of this joint-effort project (UT-Austin, UIUC, Oregon State University, University of Arkansas, and the Institute for Future Intelligence) are to address three research gaps: design cognition research to define GDT, education research to create/test materials for teaching GD/GDT, and technology development to support GD research. In design education, the development of curriculum to prepare next-generation students for effective industry use of generative AI requires a clear definition of GDT which describes the relevant cognitive processes and how to best teach them. In design research, this definition will aid the development of AI-agents for augmenting previously human-driven tasks, e.g., by providing inspiration during conceptual design to overcome fixation.

Project outcomes fall within these three gaps. First, a systematic review of literature on design thinking/design cognition topics (e.g., systems thinking, computational thinking, design creativity) is underway to define GDT by suggesting how these cognitive processes manifest during GD, how to best teach them, and to clarify the differences between GDT and other “design thinking” concepts. We have developed an Evolving Design Thinking (EDT) model to guide our review and represent how evolving technologies like generative AI drive the evolution of design thinking/cognition. Second, we have developed a text-based curriculum with CAD activities to teach basic GD principles. Student data (transcripts via think-aloud, design process and artifact data via a CAD platform) were gathered and used to refine the curriculum and explore students’ thought processes during GD. Third, software development includes the creation and continued support of Aladdin, a web-based open-source CAD/CAE platform which enables research and the creation of educational modules in renewable energy design contexts. Finally, project outcomes have been disseminated to collaborative academic (Lehigh University, Utah Tech, Hawaii Kapiolani Community College) and industrial partners (PTC Inc.) for broader impacts.