AI 賦能個人化物理模擬：打造「每位教師的專屬 PhET」
講者/Speaker： 江長屹, CHIANG CHANG-YI
本教學應用以「AI 賦能個人化物理模擬」為核心，旨在運用生成式人工智慧降低物理模擬教材的開發門檻，協助教師快速建置符合課堂需求、考題情境與學生概念迷思的互動式模擬資源。傳統物理教學常面臨抽象概念難以視覺化、理想模型與真實情境脫節、數據圖與物理過程缺乏連結等問題；而既有模擬平台雖具科學嚴謹性，卻較難因應教師即時客製化與在地化教學需求。本設計透過 AI 提示詞工程，結合 JavaScript、Three.js、P5.js 等網頁技術，使教師能以自然語言描述物理情境，快速生成具備參數調整、動態視覺化與圖表同步呈現的模擬頁面。平台已應用於斜拋運動、簡諧運動、耦合擺、卡文迪西實驗、水波槽、視深光徑、電磁感應與橡皮筋虎克定律等多種高中物理主題。相較於傳統人工撰寫程式需耗費 20 至 40 小時，AI 輔助開發可在 2 至 5 分鐘內完成原型，並能即時依照教學現場需求進行修正。此設計不僅提升教師教材設計自主性，也有助於學生透過互動操作理解抽象物理概念，進一步促進探究式學習與數位化物理教學創新。
This teaching application, titled “AI-Empowered Personalized Physics Simulations,” aims to lower the technical barriers for developing interactive physics simulations through generative artificial intelligence. Traditional physics instruction often faces challenges such as the difficulty of visualizing abstract concepts, the gap between idealized models and real-world phenomena, and students’ limited ability to connect motion processes with data graphs. Although existing simulation platforms such as PhET provide scientifically rigorous resources, they are often limited in customization and may not fully align with local curricula, specific assessment contexts, or students’ misconceptions. To address these challenges, this project integrates AI prompt engineering with web technologies such as JavaScript, Three.js, and P5.js, enabling teachers to generate interactive simulation pages through natural language descriptions. These simulations support parameter adjustment, dynamic visualization, and synchronized representations of animations and graphs. The platform has been applied to various high school physics topics, including projectile motion with air resistance, simple harmonic motion, coupled pendulums, the Cavendish experiment, water wave tank experiments, apparent depth in refraction, electromagnetic induction, and Hooke’s law using rubber bands. Compared with conventional manual coding, which may require 20 to 40 hours to develop a single simulation module, AI-assisted development can produce a working prototype within 2 to 5 minutes and allow immediate iterative refinement. This approach empowers teachers to become designers of digital learning resources while helping students better understand abstract physics concepts through interactive exploration, thereby promoting inquiry-based learning and innovation in digital physics education.
- 論文全文：AI 賦能個人化物理模擬：打造「每位教師的專屬 PhET」
- 影像： 相關影片