A Preliminary Exploration of the Disruption of a Generative AI Systems: Faculty/Staff and Student Perceptions of ChatGPT and its Capability of Completing Undergraduate Engineering Coursework
Project Overview
The document examines the role of generative AI, particularly ChatGPT, in undergraduate engineering education at Texas A&M University, focusing on its effectiveness in completing course assignments and gathering perceptions from faculty and students. It introduces the DANCE model, which offers a framework for faculty to adapt to the integration of AI in their teaching practices. The findings reveal that while generative AI presents significant benefits, such as enhancing learning and providing support for students, it also raises challenges, notably regarding academic integrity and the inconsistent performance of AI compared to human capabilities. Overall, the document underscores the necessity for careful consideration and adaptation in utilizing AI tools within educational environments, balancing innovation with ethical concerns.
Key Applications
ChatGPT for completing undergraduate engineering coursework
Context: Undergraduate engineering courses at Texas A&M University, targeting novice engineering students.
Implementation: ChatGPT was used to complete course assignments and assessments; faculty graded the outputs against student performance.
Outcomes: ChatGPT achieved passing grades in some assignments but generally performed below average compared to human students. Faculty noted its potential for personalized learning and instant feedback.
Challenges: Concerns about accuracy, inability to handle complex questions or visual data, and implications for academic dishonesty.
Implementation Barriers
Academic Integrity
Concerns regarding the potential for ChatGPT to enable academically dishonest behaviors among students.
Proposed Solutions: Developing clear guidelines on acceptable use of AI tools in coursework and enhancing faculty awareness of AI capabilities and limitations.
Performance Limitations
ChatGPT's inconsistent performance on assignments, particularly for higher-order thinking tasks and problems involving visual data.
Proposed Solutions: Implementing training for faculty and students on effective use of AI and continuously assessing AI performance in educational settings.
Project Team
Lance White
Researcher
Trini Balart
Researcher
Sara Amani
Researcher
Kristi J. Shryock
Researcher
Karan L. Watson
Researcher
Contact Information
For information about the paper, please contact the authors.
Authors: Lance White, Trini Balart, Sara Amani, Kristi J. Shryock, Karan L. Watson
Source Publication: View Original PaperLink opens in a new window
Project Contact: Dr. Jianhua Yang
LLM Model Version: gpt-4o-mini-2024-07-18
Analysis Provider: Openai