THE TASK DEVELOPMENT OF PHYSICS SUBJECT TEACHERS USING CHATGPT.
Loading...
Date
2025
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
SDU University
Abstract
This study explores the effectiveness of ChatGPT in solving of physics problems of varying difficulty levels across different topics as a generative AI language model. We mainly do aim to evaluate just how the model performs as well as integrate the model into physics education. The succeeding queries provide guidance to the research: With what accuracy does ChatGPT solve the physics problems at a variety of difficulty levels? Do certain specific physics topics vary in performance regarding amount? ChatGPT performs to a greatly better degree on low-difficulty problems, the study hypothesizes, and shows a stronger conceptual understanding rather than a numerical accuracy. Theoretical frameworks grounding the research are Technological Pedagogical Content Knowledge (TPACK), as well as the SAMR model (Substitution, Augmentation, Modification, Redefinition) used to assess the pedagogical value of AI integration in teaching. ChatGPT was in fact tested on a total of 105 physics problems, drawn from a total of seven topical areas, and a design which was quantitative and descriptivecomparative was adopted. Each response was scored through use of a structured rubric for conceptual understanding. The final accuracy was also able to be scored. Data were analyzed by descriptive statistics and t-tests. ChatGPT performs well enough on formula-based, low-difficulty problems, especially inside mechanics, as findings indicate, yet struggles upon optics and electromagnetism, abstract, complex topics. With consistency, conceptual understanding in scores happened to be higher than the accuracy of answers. The study offers several practical recommendations for integrating ChatGPT into STEM instruction. The study does also contribute in a theoretical way by applying both TPACK and also SAMR to AIassisted learning. It concludes instruction that is teacher-led should be supplemented, and not replaced, while ChatGPT can improve learning through its guided use.
Description
Keywords
Technological Pedagogical Content Knowledge (TPACK), ChatGPT, STEM instructions
Citation
Amirkhan D / THE TASK DEVELOPMENT OF PHYSICS SUBJECT TEACHERS USING CHATGPT / SDU University / Faculty of Education and Humanities