生成式AI融入「中文思辨與表達」課程之成效與挑戰：學生學習經驗分析

Effectiveness and Challenges of Integrating Generative AI Into the Chinese Critical Thinking and Expression Course: An Analysis of Student Experience

本文分析生成式AI融入國立臺灣師範大學「中文思辨與表達」（原國文）課程「自主問答」環節後學生之學習經驗。調查結果顯示，多數學生能運用AI來激盪想法、編修文句、擬定架構、協助推論，且多會先思考或寫草稿後再運用AI協作。然而，即使有AI協助，學生對「自主問答」的「文辭組織」與「問題提問」環節仍感學習困難，且不易辨識「AI生成」與「人類原創」之區別。此外，在學習過程中，當「AI生成」持續取代「人類原創」，恐會加劇上述學習困境。因此，AI若要融入教學設計，首先應向學生傳遞AI倫理觀念與運用界限，接續則是建立AI融入閱讀、思考、提問、回答的明確操作步驟，並輔以生成指令教學指引，引導學生藉實作、試誤、修正來發展個人化AI應用策略，依循「人類思考表達→AI協作生成→人機共辨→人機共修」實作步驟，實踐「思辨表達為主、AI為輔」的學習模式，以利達成學習任務。如此一來，對於提升學習成效，較能產生正面效益，且不失為培育AI運用素養之參考途徑。

This study explores the integration of generative AI into the “Self-Initiated Q&A” component of the Chinese Critical Thinking and Expression course (2024-2025). Implemented in both daytime and evening classes, the AI-assisted model emphasized human-led thinking supported by AI. Results show that students effectively used AI for idea generation, revision, and structural support, usually after initial independent reflection. Yet, persistent challenges emerged in rhetorical organization, question formulation, and differentiating human from AI work. These issues risk deepening if human input is continually replaced by AI output. To address this, the study recommends a scaffolded cycle– Human Thought → AI Generation → Human-AI Evaluation → Human-AI Co-revision– paired with explicit ethics and usage guidelines, fostering responsible AI literacy while enhancing critical thinking and expression.