Session: ASME Undergraduate Student Design Expo

Paper Number: 173742

Ai Assisted Design Processes With Emphasis on Ideation 

The use of Artificial Intelligence (AI) in the engineering design process and specifically to enhance individual design methods has been gaining traction due to its potential to create efficiencies and new capabilities. That being said, AI while can be a tremendous asset, but it can also be a detriment and occasionally is misused.  We have investigated the use of various AI tools such as Chat-GPT, Copilot, Miro Assist, Perplexity, CADscribe, Stable Diffusion, Viscom and others to evaluate their potential impact on different design methods.  The individual design methods we investigated in this research are embedded in a design process called Design Innovation (DI).  The DI process incorporates aspects of design thinking, systems design, traditional engineering design and even some business model planning.  DI is a user-centered design process that leads designers through four major steps.  Each of these four steps in the DI process incorporates a variety of different design methods.  The methods are orchestrated tasks that the design team accomplishes to complete that step in the DI process.  For example, traditional brainstorming might be used in the third step of the DI process, which is the step that focuses on ideation.  The four steps in the DI process are referred to as the “4Ds”.  The first D, Discover, entails engagement with the users or stakeholders.  This first step focuses on empathy in order to understand not only explicit but also implicit needs of the different stakeholders.  The second D, Define, uses systems engineering techniques such as functional decomposition and affinity grouping to interpret the information gathered in the Discover phase. This helps provide deeper insight into the design challenge.  Often this phase ends with the development of the core opportunity statement that drives the design work.  The third D, Develop, implements numerous ideation methods, such as traditional brainstorming. Additional methods such as mind mapping, C-Sketch rotational drawing and Design by Analogy serve to generate ideas to address the opportunity statement.  The fourth D, Deliver, incorporates strategies for prototyping and testing.  This final step also incorporates other project deliverables such as additional analysis or simulation plus design pitches and design documentation.  This current work provides an overview of our previously published research that demonstrated the effectiveness (or lack of effectiveness) of AI tools across the entire four Ds of the DI process.  This current work adds substantial new, unpublished results documenting the development and use of a tool that incorporates use of ChatGPT to produce a mind map of biological analogies to inspire innovative design solutions. The use of biological analogies to inspire innovative ideas for product development is a proven technique.  However, it can be difficult for a designer, who is often not an expert in biological systems, to know what biological systems to investigate for inspiration.  The new tool utilizes functional decomposition of a design problem, which is provided to ChatGPT as input.  We have written specific code to instruct ChatGPT to produce mind maps of these biological analogies organized intuitively via functional decomposition.  Other aspects of the biological analogies, such and the number of branches desired in the mind map and some characteristics of the biological systems can also be included in the coding provided to ChatGPT.   The mind maps typically provide dozens of biological analogies to inspire possible solutions to specific functions needed in a product design.  Designers have tested the tool and initial feedback on its effectiveness has been very positive. 

Presenting Author: Peter Velgersdyk Westmont College

Presenting Author Biography: Peter is currently an Engineering student with a Physics minor at Westmont College. He is part of an undergraduate research team that is working on the development and implementation of tools to use AI to enhance the engineering design process.

Authors:

Peter Velgersdyk Westmont College
Dan Jensen Westmont College