2025 ASEE Annual Conference & Exposition

The advent of generative AI and large language models, such as ChatGPT, has revolutionized various aspects of education. These tools have shown potential in enhancing learning outcomes by providing instant feedback, automating tedious tasks, and facilitating personalized instruction. However, their effectiveness in engineering education remains largely unexplored. This study aimed to investigate the use of ChatGPT in solving differential equations associated with electrophysiology as a pedagogical tool for senior-level biomedical engineering students.
As part of a laboratory exercise, we tasked senior-level BME students, with deriving and solving differential equations describing the kinetics of GABA receptor binding and corresponding synaptic potentials, to understand the basis for excitatory and inhibitory synapses, synaptic current, and time constant of receptor decomposition. The students worked in groups of 4 on the lab experiment, report, and exercise. The exercise consisted of students answering questions by traditional methods (hand-solving, coding, plotting, etc.) on their own. Following this, to test the efficacy of no-cost generative AI, they were asked to repeat the same set of questions, by utilizing generative AI and ChatGPT (v3.5) to generate solutions and plots The objectives of the study were to assess the feasibility of using ChatGPT for prompt engineering, compare results obtained from ChatGPT with those derived manually, and allow students to reflect on the learning experience. A secondary objective was for instructors to gain insights into student interactions with the technology and to become aware of the challenges and potential pitfalls in student learning when using generative AI for course material.
The exercise consisted of a set of questions divided into three sub-categories. (1) Deriving equations describing the kinetics of GABA receptor binding, synaptic current, and time constant of receptor decomposition. (2) Sketching plots based on given values and (3) Identify the plots in (2) either as excitatory or inhibitory. Using ChatGPT the students proceeded to generate prompts for answering questions (1) – (3). They also employed prompt engineering to create MATLAB code for generating the plots in question (2). Additionally, they asked ChatGPT to identify the type of synaptic potentials and determine the range of values for parameters required to generate excitatory and inhibitory potentials.
The results showed that while ChatGPT provided accurate solutions in some cases, its responses often differed from the manual calculations performed by the students. This discrepancy led to discussions about the limitations of AI models and the importance of verifying their results carefully. Students also reported learning about the potential biases and assumptions made during calculation, highlighting the need for critical thinking and skepticism when working with generative AI.
This study contributes to the growing body of research on the use of AI in engineering education, emphasizing the significance of pedagogical approaches that foster critical thinking, problem-solving skills, and media literacy. The findings suggest that ChatGPT can be a valuable tool for supplementing traditional instruction but must be used judiciously to avoid reinforcing misconceptions or oversimplifications.