2025 ASEE Annual Conference & Exposition

Calculus courses, typically taught in the initial years of the degree, remain a barrier to engaging and retaining students in engineering majors. These courses have long been viewed as “weeder courses,” often result in high failure rates, and have been blamed for first-year students leaving engineering majors. One reason for this impediment is a lack of connection between abstract calculus concepts and their application in students’ engineering coursework. Additionally, the problems assigned to students in these courses overlook students' prior experiences and career interests, making it harder for them to understand the connections between calculus and their chosen major. As such, there is a need to improve undergraduate engineering education by personalizing students’ calculus learning experience so that students become prepared to solve engineering problems that are complex and authentically representative of the types of problems they will face in their careers.

This project evaluates the potential of a custom AI problem-generation tool, which has been designed to bridge the gap between requisite calculus skills and the various engineering disciplines, by answering the following research question: What are the latent objectives to personalized engineering-contextualized calculus problems when generated by AI? To answer this RQ, we evaluate the AI tool ProGenie, which uniquely implements a calculus topic (e.g., partial differentiation) into an engineering domain (e.g., fluid dynamics) to rapidly create realistic and tailored problem sets for instructors. In this study, we evaluate AI-generated problems using two theoretical frameworks: Bloom’s taxonomy to assess cognitive complexity and knowledge typology. Through provisional analysis, we will identify potential strengths and limitations in the design of AI-generated problems.

We present these findings as a guide to improve the development of ProGenie and other generative AI models. Specifically, we will incorporate findings into ProGenie’s model, enhancing its capability to produce engineering-calculus problems that are sufficiently realistic, complex, and solvable given prerequisite knowledge. Problem-generating AI tools like ProGenie have the potential to strengthen students’ understanding of the connection between calculus and engineering, which creates a foundation for exploring its impact on academic success and student retention.