ChatGPT輔助翻轉教室模式下的技能學習：認知負荷與學習表現探討

Abstract

在技能學習型課程中，如何在有限的教學時間內兼顧學生的認知建構與操作實作，一直是教育現場的重要課題。隨著生成式人工智慧（AI）技術的發展，教師得以透過AI工具進行教學資源設計與備課支援，進而發展結合ChatGPT的翻轉教室教學模式，以強化學生在翻轉教室各階段的學習支持。本研究聚焦於國立臺灣師範大學「1132木器設計與製造」課程，探討ChatGPT輔助備課的翻轉教室設計對學生學習歷程與認知負荷的影響，並進一步分析不同基礎能力學生在技能學習歷程中的主觀感受與調節策略。本研究採取質性取向，透過課堂觀察與六位不同基礎能力學生之焦點訪談，深入分析學生對於三種認知負荷（本質性、外在性、增生性）的感知樣貌，及其對技能理解與實作經驗的形塑歷程。研究結果指出，ChatGPT所輔助設計的課前任務與學習資源，有助於降低學生在進入課堂前的外在性負荷，並提升課程整體的結構化支持感。學生普遍表示，動手實作與反思活動最具挑戰性，卻亦是知識深化的關鍵階段，展現增生性負荷的學習意涵；然而，基礎能力較低的學生仍須仰賴教師較多引導與資源整合，顯示本質性負荷與學習門檻密切相關。整體而言，ChatGPT輔助下的翻轉教室設計展現其在技能型課程中的應用潛力，有助於建構更有效且具回饋性的技能學習模式。In skill-based learning, balancing cognitive development and hands-on practice within limited instructional time remains a key challenge. With the rise of generative AI, tools like ChatGPT have enabled teachers to develop structured resources for flipped classroom instruction. This study, conducted in the “1132 Woodworking Design and Manufacturing” course at National Taiwan Normal University, explored how a ChatGPT-assisted flipped classroom influenced students’ learning experiences and perceived cognitive load. Adopting a qualitative approach, the research included classroom observations and focus group interviews with six students of varying prior abilities. Results showed that AI-supported pre-class materials helped reduce extraneous load and improved structured learning support. While students found hands-on and reflective tasks demanding, these fostered germane cognitive processing. However, those with lower prior knowledge experienced greater intrinsic load and relied more on teacher guidance. Overall, the integration of ChatGPT into flipped instruction demonstrated potential to enhance skill acquisition through targeted support and better cognitive load management.