Welcome to NotebookLM, an AI-powered research and writing assistant available to the Davidson College community. NotebookLM is a "source-grounded" tool that works exclusively with content you provide, such as lecture notes, research articles, project plans, or video transcripts. Unlike general-purpose AI chatbots that draw from the broader internet, NotebookLM's focus on your materials dramatically minimizes AI "hallucinations" (factual errors or fabrications), making it a reliable partner for academic work. However, no large language model is entirely free from the risk of hallucinations.
Note: This information is up to date as of November 2025.
Table of Contents
Understanding Sources: The Core of NotebookLM
Potential Use Cases at Davidson
Known Limitations and Cautions
What is NotebookLM?
NotebookLM is a source-grounded AI research assistant that works exclusively with documents you upload to a specific "notebook." Unlike general-purpose chatbots that draw from the internet, NotebookLM's responses are derived only from your materials, making it ideal for academic work requiring accuracy and direct engagement with specific sources.
Now, let's cover the simple steps to get you started.
How to Access NotebookLM
Getting started with NotebookLM is straightforward. Access is tied to a standard Google account, making the tool readily available to the entire Davidson community without any complex setup.
When signed in with your Davidson account, Google does not use your uploaded sources or chat conversations to train its AI models, ensuring your academic work remains private.
- Navigate to the website: go to notebooklm.google.com.
- Sign in: Use your Davidson account to sign in.
- Upload Documents: Once you're in, the first step is to provide the tool with the material it will work with: its sources.
CRITICAL: Chat responses are temporary and disappear when you reload the page. Click "Save to note" to preserve important content. Saved notes can be combined and converted into new sources for deeper, iterative analysis.
Understanding Sources: The Core of NotebookLM
"Sources" are the foundation of NotebookLM. The AI can only analyze and answer questions about the information you upload. Mastering source management is key to practical use.
A. Supported Source Types
NotebookLM can process a wide variety of information formats, allowing you to build a comprehensive knowledge base for any project.
- From your computer or Google Drive: PDF, Text (.txt), Markdown (.md), Google Docs, and Google Slides.
- Web Content: Web URLs (for text content only) and public YouTube video URLs (which imports the video's text transcript).
- Direct Input: Text you copy and paste directly into the tool.
- Audio Files: MP4, M4A, AIFF, AAC, CAF, AMR, WAV, MP3, and Opus. Note that this feature is experimental, and the quality of the transcription depends heavily on the clarity of the recording.
B. Source and Notebook Limits
A single notebook can hold up to 50 sources, with each source limited to 500,000 words or 200 MB, and approximately 25 million total words per notebook.
| Feature | Limit (free plan) |
| Max sources per notebook | 50 |
| Max size per source | 500,000 words or 200 MB |
| Total words per notebook | Approx. 25 million |
| Max number of notebooks | 100 |
Key Features and Capabilities
Once your sources are uploaded, NotebookLM offers tools to analyze texts, synthesize ideas, and test your knowledge.
A. Interactive Chat and Source-Grounded Answers
Ask specific questions about your uploaded documents through the chat interface. Responses include inline citations that link directly to the source material, allowing you to verify accuracy and context instantly.
B. Automatic Content Generation
NotebookLM generates functional study materials, including summaries, study guides, timelines, briefing docs, FAQs, and custom reports. These automated documents help you quickly digest and organize information from your sources.
C. AI-Generated Podcasts (Audio Overviews)
Generate podcast-style discussions between two AI voices that explore your sources conversationally. An interactive mode lets you pause and ask follow-up questions, with the discussion adapting in real time based on your materials.
D. Mind Map Visualization
Automatically creates interactive diagrams showing how concepts in your sources connect and relate. Particularly useful for organizing complex research topics or preparing outlines for essays and presentations.
E. Video Generation (AI-Generated Visual Summaries)
Generate short video summaries with synthesized narration and on-screen visuals. Videos can be downloaded or embedded, making them ideal for presentations or sharing with collaborators.
F. Flashcards and Quizzes
Automatically creates flashcards and quizzes from your materials to support active learning and test comprehension.
Potential Use Cases at Davidson
NotebookLM is a highly adaptable tool for academic life. Here are a few ways it can be applied within the Davidson community.
- Academic Research and Study: If permitted by professors, upload lecture notes, textbook chapters, and academic PDFs into a single notebook. Use the Study Guide generator to test knowledge, get summaries of dense readings, or quickly find answers across all materials.
- Project Management and Synthesis: Student groups can upload project plans, meeting transcripts, and background articles to create briefing docs for supervisors, FAQs for new members, or timelines of key milestones. The Audio Overview feature can quickly bring new team members up to speed.
- Complex Topic Analysis: Upload multiple sources on a complex topic—such as quarterly earnings reports or philosophical texts. To compare arguments, identify trends, and synthesize key takeaways into a single, cohesive note.
Known Limitations and Cautions
NotebookLM is a powerful tool, but understanding its limitations ensures you get the most out of it.
- Source Quality is Paramount: NotebookLM's output depends entirely on the quality of uploaded sources. Well-structured, accurate documents yield insightful results, while low-quality sources produce unreliable outputs.
- Not Optimized for Creativity: NotebookLM is designed for accuracy, analysis, and synthesis of existing information, not creative writing or brainstorming novel ideas.
- No Conversational Memory: Each chat interaction is stateless and does not retain information from previous questions unless it's present in uploaded sources or saved in a note.
- Outputs Must Be Saved Manually: Chat responses are ephemeral and will be lost permanently if you close or reload the notebook without clicking "Save to Note."