使用AI生成圖像協助學生語文創作之成效：AI自我效能、學習興趣、求知好奇心、AI使用挫折感與靈光乍現之相關研究

Abstract

隨著AI技術快速發展，教育現場迎來新契機，語文教育亦不例外，透過生成式AI圖像工具輔助語文創作，可視覺化學生抽象想像，激發創意靈感，並促進語文表達的深度與廣度。本研究透過使用生成式AI工具輔助學生學習，探討AI生成圖像學習之「AI自我效能」、「學習興趣」、「求知好奇心」、「AI使用挫折感」、「靈光乍現」與語文創作表現之關係。本研究採用實驗研究法及問卷調查法，以新北市國小高年級學生為主要研究對象，在語文課程中導入Microsoft Copilot生成式AI為教學工具，協助學生創作圖像及應用於語文寫作。學生於教學實驗前、中、後填寫使用AI圖像與語文創作問卷量表，並在課程結束後收集與分析128名有效參與者的數據，探討生成式AI圖像對學習歷程的支持效果與潛在關係。本研究結果顯示，AI自我效能對學生的求知好奇心具有顯著正向影響，興趣型好奇心正向影響學生的學習興趣，並對AI使用挫折感呈現負向關係。學習興趣有助於提升學生在語文學習過程中的靈光乍現經驗，而AI使用挫折感則會對靈光乍現產生負向影響。最後，靈光乍現對語文學習成績表現有正向影響。此外，不同年級會影響AI使用經驗與創造表現，顯示教育現場的AI應用需顧及學習背景與資源差異，強調師資AI素養與學生操作支持的重要性。本研究發現生成式AI圖像有助於啟動學生靈感與創造力，提升語文創作動機與成效，並為課程設計、數位教學政策與AI教育應用提供具體建議與實證依據。With the rapid advancement of AI technology, new opportunities have emerged in education, including language instruction. By incorporating generative AI image tools into language composition, students’ abstract imaginations can be visualized, thereby stimulating creative inspiration and enriching both the depth and breadth of language expression. This study investigates the relationships among AI self-efficacy, learning interest, epistemic curiosity, AI-related frustration, Aha moments, and students’ language performance when supported by generative AI tools.Employing both experimental and survey methods, this study targeted upper-grade elementary students in New Taipei City. Microsoft Copilot was integrated into language courses to assist students in generating images and applying them in writing. Students completed questionnaires on their AI-assisted composition experiences before, during, and after the intervention. Data from 128 valid participants were collected and analyzed to examine the supportive effects of AI-generated images and their interrelations.The findings revealed that AI self-efficacy significantly enhanced students’ epistemic curiosity. Interest-type curiosity positively influenced learning interest and negatively predicted AI-related frustration. Learning interest promoted Aha moments, while frustration hindered them. Aha moments, in turn, positively influenced students’ language performance. Additionally, variations in grade levels affected students’ AI experiences and creativity, highlighting the need to consider learner backgrounds and resource disparities, and emphasizing the importance of teacher AI literacy and student support.This study concludes that generative AI image tools can effectively activate students' inspiration and creativity, enhance motivation and performance in language composition, and offer concrete implications for curriculum design, digital teaching policies, and AI integration in education.