TITLE:
Using ChatGPT to Develop University Instructors’ TPACK and Students’ Self-Regulated Learning
AUTHORS:
Syh-Jong Jang
KEYWORDS:
ChatGPT, University Flipped Classrooms, TPACK, Self-Regulated Learning, Mixed Analysis
JOURNAL NAME:
Open Journal of Social Sciences,
Vol.13 No.5,
May
30,
2025
ABSTRACT: ChatGPT has gained popularity on social media and proves to be a valuable resource in higher education, assisting students in understanding and evaluating topics by offering topic-specific information, suggesting undiscovered aspects, and introducing new research topics. However, there are currently limited studies on ChatGPT generating text and creating scientific teaching resources in the context of science education. Moreover, there is no research on how university students or teachers apply it to influence teachers’ teaching or students’ learning ability. The study aimed to explore how a university instructor used the ChatGPT flipped model to impact university students’ perceptions of the instructor’s TPACK and students’ self-regulated learning (SRL) ability in a general education course. The results show that while the total mean values of TPACK pre- and post-tests did not reach significant differences, the self-regulated learning (SRL) outcomes of university students obtain significant differences. The research indicates that ChatGPT can assist the university instructor in sourcing text materials, facilitating class discussions, and providing individualized guidance. Additionally, ChatGPT helps the instructor find various topics and offers rich and diverse cross-domain content suggestions. ChatGPT facilitates quick access to required information and reducing search time. It also assists students in finding various topics, offering suggestions, explanations, and aiding in task organization. While it offers different types of subject information and aids in task completion, students should self-evaluate the text information provided by ChatGPT to ensure correctness and suitability. Despite functioning as an artificial intelligence environment, it lacks the ability to empathize with users. The research implications of this study are presented along with relevant suggestions.