Abstract:

The integration of artificial intelligence (AI) in education has evolved toward specialized environments that enable personalized learning experiences using proprietary sources. This study analyzes two Google platforms, NotebookLM and Gemini, and their potential in academic research, instructional material creation, assessment, and teacher management.

NotebookLM organizes information into thematic “notebooks” that can contain up to 50 documents and 500,000 words, supporting PDF, text, Markdown, MP3 audio, integrations with Google Drive, and the analysis of web pages or videos through transcriptions. Its focus on privacy and academic rigor ensures that responses are based solely on the provided sources, including direct references to minimize errors or model hallucinations. NotebookLM’s “Studio” feature transforms these sources into educational resources such as audio or video summaries, interactive mind maps, study guides, flashcards, infographics, and presentations. These materials can be adapted by educational level, language, tone, and pedagogical objective, enabling the creation of personalized and accessible content for diverse learning contexts. The platform also facilitates the generation of materials tailored to different curricular levels and bilingual glossaries through structured prompt engineering, which combines expert roles, educational context, selected sources, examples, and desired output formats.

Gemini stands out for its multimodal versatility and automation capabilities. It allows the creation of personalized assistants (“Gems”) configured with specific instructions and reference documents to act as experts in rubrics, planning, FAQs, or pedagogical advising. Its “Deep Research” function performs comprehensive web searches and generates detailed technical reports.

Other tools are also discussed in this article, such as the “Canvas” environment, which supports the editing of extensive documents by adjusting tone and length, and enables the generation of outlines, quizzes, infographics, summaries, and materials exportable to Google Docs. Canvas’s “Artifacts” allow the creation of interactive applications through code, such as educational simulators or dynamic visualizations. In the assessment domain, AI can automatically generate open and closed tests, multiple-choice quizzes, true/false items with explanations, or fill-in-the-blank exercises. These resources can be integrated into Google Forms, Kahoot, or Plickers, and facilitate the design of rubrics, weighted checklists, and co-assessment tools adapted for students.

Finally, despite technological advances, limitations remain in producing precise technical images and managing copyright. It is emphasized that AI functions as a teaching assistant, with the educator maintaining ultimate responsibility for reviewing and validating generated content.

Keywords:

Educational technology tools, NotebookLM, Gemini AI, teacher support.