Core Concepts
The authors aim to evaluate the effectiveness of ChatGPT in completing engineering coursework and explore faculty and student perceptions. They utilize the DANCE model to understand the impact of disruptive tools like ChatGPT on the learning environment.
Abstract
A study was conducted to assess OpenAI's ChatGPT tool for engineering coursework at Texas A&M University. The research aimed to understand perceptions, analyze survey data, and evaluate ChatGPT's performance across various engineering courses. Findings suggest that while ChatGPT excelled at basic tasks, it struggled with more complex assignments, raising concerns about academic integrity and educational implications.
The study involved surveys distributed to students, faculty, and staff regarding their perceptions of ChatGPT's impact on academia. Results indicated mixed views on academic dishonesty facilitated by ChatGPT. Faculty expressed discomfort with students using external resources like ChatGPT for coursework. Network analysis revealed distinct groups of faculty attitudes towards GAI systems.
Performance assessments showed that while ChatGPT performed adequately in lower-level courses, it struggled in higher-level engineering coursework. The AI system often provided general answers lacking depth or accuracy required for passing grades. Future research aims to refine teacher development programs integrating emerging technologies like GAI systems.
Overall, the study sheds light on the evolving role of generative AI in education and emphasizes the need for continuous adaptation to technological disruptions in academia.
Stats
835 student responses and 248 faculty/staff responses were collected.
An alpha value of 0.036 indicated students were more familiar with ChatGPT than faculty.
An alpha value of 0.427 showed no significant difference between student and faculty perceptions of academically dishonest behavior due to ChatGPT.
A paired t-test with an alpha of 9.14E-14 revealed a significant increase in perceived likelihood of honor code violations post-ChatGPT release.
Faculty comfortability with students using ChatGPT varied, with a bimodal distribution observed.
Quotes
"Understanding how generative AI will impact engineering education is far from completely understood."
"ChatGPT excelled at basic tasks but struggled with more complex assignments."
"Faculty expressed discomfort with students using external resources like ChatGPT for coursework."