學術文獻的快速增長使研究者面臨文獻整合困難的挑戰，但傳統文獻管理工具難以有效支援跨文獻的語意整合與知識提煉。而Google於2023年推出的NotebookLM，以研究者自行上傳的文件作為封閉式知識庫，採「以來源為基礎的推理」機制，在保留生成式AI強大語意分析能力的同時，可大幅降低學術引用中的幻覺風險。本文旨在系統性介紹NotebookLM在教育研究文獻整理上的核心功能，並以社會認知生涯理論（Social Cognitive Career Theory, SCCT）量化縱貫研究文獻為示範材料，逐步展示文獻匯入、精準問答、跨文獻比較與結構化摘要等操作流程，探討其對教育研究者建立文獻管理工作流程的應用價值，同時說明工具的使用限制與學術倫理注意事項。

The rapid growth of academic publications has made literature integration increasingly challenging for researchers, and traditional reference management tools are limited in supporting cross-document semantic synthesis and knowledge refinement. NotebookLM, launched by Google in 2023, addresses this gap by treating user-uploaded documents as a closed knowledge base and adopting a source-grounded reasoning mechanism, which preserves the semantic integration capabilities of generative AI while substantially mitigating the risk of hallucinated citations in academic work. This article systematically introduces NotebookLM’s core functions for literature management in educational research, using twelve quantitative longitudinal studies grounded in Social Cognitive Career Theory (SCCT) as demonstration material. The article walks through the workflow of source importing, source-fidelity verification through hallucination-trap testing, precise question-answering, automated cross-document briefing, mind mapping, and structured comparison via Data Tables. It also discusses the application value of these functions in helping educational researchers establish efficient literature-management workflows, while addressing the tool’s current limitations and the ethical considerations of AI use in academic research.