Development of ChatGPT-Based Physics Simulations: From the Doppler Effect to Electric Field Exploration
Tia Jannah Tertia*, Ghonia Millata Mallia Shofa, Meita Puteri Handayani,Sparisoma Viridi

Department of Physics,Bandung Institute of Technology
Jalan Ganesha 10, Bandung 40132, Indonesia
*tertiatia12[at]gmail.com

Abstract

The utilisation of Large Language Models (LLMs) such as ChatGPT offers educators opportunities to independently develop digital learning media without advanced programming expertise.This research evaluates the robustness of the iterative prompt engineering method by transitioning from Doppler effect simulations to electric field concepts, which entail higher vector complexity.Whilst this method was previously successful for the Doppler effect, its implementation for electric fields revealed technical challenges, specifically visual inaccuracies in field line rendering.These findings suggest a performance threshold for AI in automatically managing complex interactive visualisations without manual code intervention.The study concludes that although prompt engineering significantly accelerates the development of simulation frameworks,the educator^s role as a technical evaluator remains crucial to ensure the scientific validity and final quality of the resulting educational media.

Keywords: physics simulation, ChatGPT, Doppler effect, electric field, prompt engineering, physics learning

Topic: Physics Education