Art & Design- Instructor- UC Berkeley Extension
Brazosport College
Lake Jackson, Texas
Title
Leveraging Custom GPT Models to Enhance Instruction in Higher Education
Synopsis
This study examined the integration of an instructor-grounded custom GPT (Generative Pre-Trained Transformer) model, “AutoCAD Studio@UC Berkeley,” within a live online AutoCAD course in higher education workforce programming. The purpose of the intervention was to provide on-demand, discipline specific instructional support aligned with course content, drafting standards, and instructor workflows while promoting more efficient use of valuable faculty time. A mixed methods design was used to compare pre and post intervention student evaluation data and analyze structured student reflections regarding GPT- assisted learning interactions. Quantitative findings indicated that students perceived the custom GPT as highly effective for procedural clarification, troubleshooting, and reinforcement of drafting standards. Findings align with emerging literature suggesting that bounded, discipline specific AI systems may improve learner support while reinforcing faculty guided instruction.