Conference Session

Design in Engineering Education Division (DEED) - AI and Digital Futures in Design Education

Collection

2025 ASEE Annual Conference & Exposition

Authors

Daniel D. Jensen, Westmont College; William Johnston Allison; Camila Rapalo; Mark Rogers; Gregory Reich, Air Force Research Laboratory, Aerospace Systems Directorate; Landon Thomas Vanderhyde

Tagged Divisions

Design in Engineering Education Division (DEED)

content across a broad spectrum (e.g., texts, images, orprogramming code) for various domains based on basic user prompts” (13). There are a varietyof AI-T that have GenAI capabilities. LLMs such as ChatGPT, OpenAI, and Google Bard can beused to create unique natural language texts for research paper summaries or outlines forexample (8). Meneske explains, “Midjourney and DeepBrain AI are diffusion models that cancreate diagrams (e.g., concept maps), images, and videos from textual or visual inputs.Engineering education, in particular, can benefit from integrating and utilizing generative AItechnologies to improve instructional resources, develop new technology-enhanced learningenvironments, reduce instructors’ workloads, and provide students with

Conference Session

Design in Engineering Education Division (DEED) - AI and Digital Futures in Design Education

Collection

2025 ASEE Annual Conference & Exposition

Authors

Daniene Byrne Ph.D., Stony Brook University

Tagged Topics

Diversity

Tagged Divisions

Design in Engineering Education Division (DEED)

ofcomputing but nearly every field of science and human endeavor[5]”. Some in the industry haveframed them as the first steps toward Artificial General Intelligence (AGI), meaning systems thatthink more like humans in numerous ways. Like humans, AGI will have the ability to ‘think’about many things across many domains, requiring different recall of datasets and intuition.This literature survey describes how policies around responsible governance are taking shape asstrong AI technologies emerge, and public interaction with them expands exponentially. InNovember of 2022, the first generative AI (GenAI) ChatGPT, created by OpenAI, was widelyreleased to the public. Earlier versions had been in development and were tested and used foryears but the public

Conference Session

Design in Engineering Education Division (DEED) - Emerging and Sustainable Design Practices

Collection

2025 ASEE Annual Conference & Exposition

Authors

Russell K. Marzette Jr., The Ohio State University; Bhavana Kotla, The Ohio State University; Cal King, The Ohio State University

Tagged Divisions

Design in Engineering Education Division (DEED)

method to find survivors that have access to their phones and can connect to an automatedinternet source. Second, it's useful in hazardous situations like the Turkey earthquake, likementioned before. Third, it offers quick response to people, minimized lives blasting theaftermath.”Sub-theme 4: NoveltyThis sub-theme highlights participants acknowledging the novelty and uniqueness of their createdsolutions (e.g., how the product stands out).“Our project, Flowware, stands out by using ChatGPT API to offer smart, personalized financialmanagement while displaying your finances through react flow, creating a dynamic, real-timemap of your money.”Theme 2: Design and ApplicationThe design and application theme includes challenges, design considerations

Conference Session

Design in Engineering Education Division (DEED) - Best in DEED

Collection

2025 ASEE Annual Conference & Exposition

Authors

L'Nard E.T. Tufts II, Stanford University; Alessandra O. Napoli, Stanford University; Shima Salehi, Stanford University; Anna Lisa Boslough, Stanford University

Tagged Divisions

Design in Engineering Education Division (DEED)

efficacy.AcknowledgmentsWe utilized resources from Stanford University's "AI Playground" to explore and validate ourapproaches to incorporating AI tools into the feedback generation process. Through this portal,Anthropic's Claude.ai, version 3.5-Sonnet, helped automate the analysis of student reflectiveresponses by identifying general themes, common omissions, unique realizations, and evidenceof reflective practice. OpenAI's ChatGPT, version 4o, helped generate reflection scores forstudent responses, providing a quantitative measure of reflective practice. We thank thedevelopers of these tools which are available using the links below.https://aiplayground-prod.stanford.eduhttps://claude.aihttps://chat.openai.comReferences[1] T. Anderson and J. Shattuck, "Design

Conference Session

Design in Engineering Education Division (DEED) - Innovative Assessment Strategies in Design

Collection

2025 ASEE Annual Conference & Exposition

Authors

Jessie Cortez, Texas A&M University; Joanna Tsenn, Texas A&M University

Tagged Divisions

Design in Engineering Education Division (DEED)

sampleBackground sections responding to the following problem statement: The McDonnell Douglas DC-10's outward-opening cargo door has a faulty locking mechanism that, upon failure, causes the door to open and the plane to explosively decompress. The project sponsor, a representative of McDonnell Douglas, has asked the design team to redesign the aft cargo door to prevent accidental opening during flight.Each background section was written with the aid of ChatGPT to simulate problems withrhetorical appropriateness and formatting and organization observed in students’ backgroundresearch sections in previous semesters. For example, the technical writing coordinator promptedChatGPT to write a background research section focused on types of cargo doors

Conference Session

Design in Engineering Education Division (DEED) - Team-Based and Experiential Learning

Collection

2025 ASEE Annual Conference & Exposition

Authors

Prarthona Paul, University of Toronto; Anipreet Chowdhury, University of Toronto; Loura Elshaer, University of Toronto; Anushka Sethi, University of Toronto; Hamid S Timorabadi P.Eng., University of Toronto

Tagged Divisions

Design in Engineering Education Division (DEED)

project-based courses. Theexisting pre-trained models did not yield good enough results; therefore, we decided to train ourown. We extracted sample tasks from 200 syllabi from engineering project-based courses. Someof these are publicly available syllabi, from real engineering courses from different NorthAmerican Universities, while others are of fictional engineering courses developed by generativeLLM tools, such as ChatGPT and Microsoft Copilot based on the formats of the real syllabi.These extracted tasks were then labelled with their corresponding classes, which were used totrain a RoBERTa model. This model performed better than the pre-trained models, as it had anF1 score closer to the requirements for the project (outlined in Appendix A