Don’t fret; Google NotebookLM has your back 📚

 

Since the launch of Notebook Language Model (NotebookLM) by Google across 200 countries on June 6th, I have been dabbling in this Generative Artificial Intelligence (GenAI) assisted note-taking browser-based application. Initially, I thought it was yet another glorified note-taking app, but I was pleasantly surprised by its enhanced capabilities, which may prove extremely useful, especially for academia, students, journalists, or knowledge workers in general.

In this article, I'll explore the potential of NotebookLM in terms of 1. Productivity (to curate and retrieve), 2. Discovery (to integrate and synthesise), and 3. Collaboration (to share and cross-pollinate). Moreover, I'll share my experience in my personal and work-related use cases.

In Diagram 1, you can create an individual Notebook of a specific theme. Each Notebook allows up to 50 sources of information in different formats, namely Google Drive, PDF, TXT, URL, and Copied Text. Each source can contain up to 500,000 words. Each Notebook is allowed up to 1,000 accompanying Notes.

Diagram 1: Schematic Diagram of NotebookLM

I'm particularly fond of the "Source Grounding" feature. This ensures that all insights and summaries generated by NotebookLM are directly linked to the specific sources I've provided. This is crucial for maintaining academic rigor and ensuring the traceability of all claims. More importantly, as claimed by Google, these ground sources will not be used for training their language model.

 1. Productivity (to curate and retrieve)

Like most note-taking apps, NotebookLM stores, organises, and retrieves your database of materials of different formats in an instant. Unlike other note-taking apps which rely on "Tagging" or assigning keywords to different pieces of information (which can be laborious, unintuitive, and too constricting to the assigned label, leaving no room for cross-referencing), NotebookLM simply lets GenAI (Gemini 1.5 at the time of this writing) do its stuff.

2. Discovery (to integrate and synthesise)

For instance, I once asked NotebookLM to find connections between seemingly unrelated concepts in my research, and it provided insights that I might have otherwise missed. As an academic writer, I'm often tasked with synthesising information from numerous sources, identifying key themes, and drawing connections that might not be immediately apparent. NotebookLM excels in this area. Its "Notebook Guide" feature, for instance, can transform a collection of sources into a structured outline, complete with FAQs and key insights. This is an invaluable tool for quickly getting to grips with a new body of research. Perhaps most exciting, however, is NotebookLM's ability to suggest new ideas and connections. By analysing the sources I've provided, the AI can propose novel perspectives or highlight areas that warrant further exploration.

One promising application is the facilitation of interdisciplinary research. NotebookLM's ability to integrate and synthesise information from diverse sources allows researchers to draw connections across fields that might otherwise remain siloed.

 

3. Collaboration (to share and cross-pollinate)

Furthermore, academics engaged in collaborative projects will benefit from NotebookLM's enhanced collaboration features. Teams can share notebooks containing their respective research findings, which the AI can then analyse to identify complementary insights or conflicting data points. This facilitates more robust peer reviews and constructive feedback loops, ultimately enriching the quality of the research output. The AI can then generate a collective study guide that amalgamates everyone's inputs, ensuring comprehensive coverage of the subject matter while accommodating different perspectives and learning styles.

 My Work/Personal Use Cases:

 1. Work (A Research Paper)

As part of a collaborative research project, I've included seven interview transcripts and tables displaying questionnaire results. The AI's responses will feature numbered links. Clicking on these links will reveal the specific, relevant text excerpts from the source documents that inform the AI's answer to your query. The responses are spot on.

 My prompts:

1.     Find the common perceptions of these interviewees with regards to GenAI in their learning and teaching. In your responses, name the interviewees as well as their quotes in verbatim.

2.     Find the contradictory statements among these interviewees with regards to GenAI in their learning and teaching. In your responses, name the interviewees as well as their quotes in verbatim.

3.     Find any interesting and insightful comments by these interviewees with regards to GenAi. In your responses, name the interviewees as well as their quotes in verbatim.

4.     What are the limitations of this study? Can you suggest 3 research gaps and 3 possible future research to address the limitations and/or the research gaps?

 

2. Personal (Books Read and Quotes Highlighted)

I’ve been an avid reader in the past (not so much nowadays). As a committed Amazon Kindle reader, I’ve amassed a sizable quotes collection, which I found insightful and interesting, from the hundreds of books I’ve read. I managed to export this information into NotebookLM. One advantage of using NotebookLM is that the export format need not be rigid as GenAI is able to figure out the details.

 My prompts:

1.     These are all the quotes I've highlighted from the books I've read over many years. Please speculate and discern my interests and sensibility as well as what you can tell about me.

2.     Can you select some of the most interesting quotes that are contradictory between the different authors? Please include the citations verbatim as well as the specific author and book.

 

3. Personal (Personal Essays and Reflections)

I managed to upload most of my writings that exist in digital format (except of course those which were handwritten). 

1.     Given the advent of GenAI, what are the possible impact of bad actors on democracy, human rights and freedom of speech?

2.     These are some of my writings over many years. Please speculate and discern my interests and sensibility as well as what you can tell about me.

As an academic, I have to admit that while Google NotebookLM has many advantages, it is not without its potential drawbacks. One major concern is the risk of over-reliance on AI-generated content. As authors, we must remain vigilant to ensure that our work retains its originality and authenticity. There's a risk that overuse of NotebookLM could lead to a homogenisation of academic writing styles, potentially stifling the diverse voices that are crucial to scholarly discourse.

Another potential drawback is the issue of privacy. Although NotebookLM operates as a closed system, uploading sensitive research materials to any cloud-based platform raises questions about data security and intellectual property rights. Researchers working on confidential or proprietary information may need to exercise caution when using such tools.

In addition, Google is notorious for launching many of its products with great fanfare and then unceremoniously 'killing' them. I wouldn't be surprised if NotebookLM suffered the same fate.

Despite these concerns, I believe that NotebookLM, if used judiciously, can greatly enhance our productivity and creativity as writers.

Lastly, I prompted NotebookLM that “I'm asked to write an article on the usage of Google NotebookLM and how it can revive memories which otherwise left tucked away. Can you find something which I've written on this matter before.”

To end this article, retrieving a paragraph of a fictional story I wrote more than twenty years back, “I have long come to believe that there is something magical about diary. They have this power of reliving memories which otherwise left in the recesses of the mind, forever lost, forever forgotten…All it takes is a key to unlock them. However this key is no longer a diary but NotebookLM.”

To watch a demo using NotebookLM by the author: