| 摘要: | 隨著擴增實境與人工智慧技術的快速發展，其在科學教育中的應用日益普遍，特別是在協助學生理解如光學等抽象概念上展現顯著成效。本研究旨在建置一套結合HoloLens 2擴增實境頭戴式裝置、實體光學儀器與AI對話系統（ChatGPT與Whisper）的沉浸式學習環境，探討不同學習情境（個別與合作）與AI輔助對國中生在「凸透鏡成像」科學主題上的學習成效、認知負荷與學習動機的影響。研究對象為92名國中一年級學生，依學習條件分為三組：單人無AI組、單人有AI組與合作有AI組，皆使用HoloLens 2進行實驗操作。AI輔助組可透過語音向系統提問，並獲得即時引導與回饋。研究透過光學概念學習單、光學概念測驗前後測、認知負荷與學習動機問卷等工具進行資料蒐集，並輔以錄影與系統互動記錄分析學生的學習歷程。結果顯示，合作有AI組在學習成效上表現最佳，特別在高層次概念應用與遷移學習方面顯著優於其他組別；AI輔助亦能有效提升學生的觀察與概念理解能力。在認知負荷部分，AI輔助組學生在內在認知負荷顯著低於無AI組，顯示系統能有效減輕學生的學習壓力；而在內在動機與自我效能方面則未達顯著差異。錄影分析指出，擴增實境結合AI的學習環境能促進學生釐清科學概念與互動討論，但仍存在資訊不同步與合作品質不一等挑戰。整體而言，本研究建構的AR+AI科學探究系統能有效提升學生對光學概念的理解與應用能力，並促進主動學習與合作探究。研究成果可作為未來發展沉浸式教育科技與設計合作學習活動的重要參考。;With the rapid development of augmented reality (AR) and artificial intelligence (AI) technologies, their application in science education has become increasingly widespread, particularly in assisting students in understanding abstract concepts such as optics. This study aims to construct an immersive learning environment that integrates the HoloLens 2 AR headset, physical optical instruments, and AI dialogue systems (ChatGPT and Whisper). The study investigates the effects of different learning scenarios (individual and collaborative) and AI assistance on junior high school students’ learning outcomes, cognitive load, and learning motivation in the science topic of “image formation by convex lenses.” The participants were 92 seventh-grade students, divided into three groups based on learning conditions: an individual group without AI, an individual group with AI, and a collaborative group with AI. All groups conducted experimental operations using the HoloLens 2. Students in the AI-assisted groups could ask questions via voice and receive real-time guidance and feedback. Data were collected through an optics concept worksheet, pre- and post-tests on optics concepts, cognitive load and learning motivation questionnaires, along with video recordings and system interaction logs to analyze students’ learning processes. The results showed that the collaborative AI-assisted group achieved the best learning performance, particularly excelling in higher-order concept application and transfer learning compared to the other groups. AI assistance also effectively enhanced students’ observation and conceptual understanding skills. Regarding cognitive load, students in the AI-assisted groups experienced significantly lower intrinsic cognitive load than those without AI, indicating that the system effectively reduced learning pressure. However, no significant differences were found in intrinsic motivation or self-efficacy. Video analysis revealed that the AR+AI learning environment helped students clarify scientific concepts and promoted interactive discussions, though challenges such as information asynchrony and varying collaboration quality remained. Overall, the AR+AI science inquiry system developed in this study effectively enhanced students’ understanding and application of optical concepts while fostering active learning and collaborative inquiry. The findings provide valuable insights for the future development of immersive educational technologies and the design of collaborative learning activities. |
