Developing a ChatGPT-Based Tool for Physics Experiment Teaching

Abstract

This paper examines how advanced AI assistants can help physics educators create practical teaching tools without specialized programming skills. Using the square-wave synthesis experiment as a case, we target common obstacles in laboratory instruction-complex setup, unstable signals, and limited class time-and show how AI-assisted development can shift attention from wiring and calibration to core physical ideas. To address this need, we guided an AI assistant through iterative prompts to generate a clean, runnable program that visualizes square-wave synthesis from its component sine waves. The tool supports step-by-step construction of the waveform, adjustable parameters (amplitude, frequency, and phase), and immediate comparison with an ideal reference using simple error measures. We packaged the result as a standalone application so it runs reliably on standard classroom computers, enabling pre-lab demonstrations and interactive exploration that reduce procedural friction and highlight underlying concepts. Building on this proof of concept, we argue the same workflow extends to other topics-such as simple harmonic motion and optical interference-where adjustable parameters and real-time visualization deepen understanding. We conclude that AI-assisted co-design can improve teaching efficiency, enhance student engagement with foundational principles, and provide a scalable path for developing customizable physics education tools.