探索生成式AI助理在師生聊書中的潛力

Exploring the Potential of Generative AI Assistants in Teacher-Student Book Talks

雖然聊書是一種與他人討論及交流書籍內容或閱讀心得極具潛力的活動，然而，在學校推動仍有其困難。因此，本研究探討生成式AI助理在師生聊書活動中的潛力，旨在解決教師於進行聊書活動時的高教學負擔，並提升學生的參與度及學習成效。本研究設計並開發了一款「AI助理」，該系統透過自動記錄和分析師生聊書內容，在「示範與鷹架」、「引導學生參與」及「轉移責任」三個教學面向上為教師提供輔助，以提供個別化的指導建議，減少教師在記錄、整理與分析上的時間投入。本研究實施於桃園市某國小的二至六年級的五位教師及75位學生，為期一學期。並邀請五位教師分別參與三次訪談，共蒐集75段錄音並分析。研究發現，AI助理不僅能夠幫助教師簡化閱讀活動的管理，以減輕工作量，還能將教師的教學工作從原先記錄、整理和分析這類低層次且繁瑣的工作，轉變成為個別學生提供專業的教導與指引的高層次教學工作。此外，研究建議進一步擴充AI助理的分析功能，讓教師能夠根據自身需求選擇分析內容，並透過文字互動更深入地理解學生的學習狀況。本研究證明了AI助理在閱讀活動中可作為有效的教學輔助工具，並具有顯著的潛力可促進師生互動與教學成效。

Although book talks, where individuals engage in conversations about book content or reading reflections, have great potential, there are challenges in promoting such activities in schools. Therefore, this study explores the application and potential of generative AI assistants in teacher-student book talks, aiming to reduce the heavy teaching burden on teachers and enhance student engagement and learning outcomes. The study designed and developed an “AI Assistant” system, which automatically records and analyzes teacher-student book talk content, providing support in three teaching dimensions: “Modeling and Scaffolding,” “Guiding Student Participation,” and “Shifting Responsibility.” The system offers personalized guidance and recommendations, reducing teachers’ time recording, organizing, and analyzing. This study was conducted with five teachers and 75 students from grades two to six in an elementary school in Taoyuan City over one semester. The five teachers participated in three interviews, resulting in 75 recorded transcripts for analysis. The findings indicate that the AI assistant not only helps simplify the management of reading activities, reducing the workload but also shifts teachers’ efforts from low-level tasks, such as recording and organizing, to high-level tasks, such as providing professional guidance and personalized instruction. Additionally, the study recommends further expanding the AI assistant’s analysis functions, allowing teachers to select analysis content based on their needs and enabling a more in-depth understanding of students’ learning progress through text-based interaction. This study demonstrates that AI assistants can be practical tools in reading activities and have significant potential to enhance teacher-student interaction and overall teaching effectiveness.