Your Trusted AI Assistant Guide

The AI Privacy Dilemma

If you’ve been keeping up with our latest articles, you’ll recognise our cautious stance regarding how AI companies manage your data. The current landscape resembles a lawless frontier where intellectual property is frequently overlooked, owing to issues like the aggressive scraping of sites such as URTech.ca and Commodore.ca.

Many AI platforms, including ChatGPT, Claude, and Meta AI, essentially consume the information you provide. Uploading sensitive business presentations or academic research can lead to your confidential material becoming part of their vast training datasets. This poses significant risks for professionals and students alike; proprietary information and original research should never be placed in a system where its fate is uncertain.

How Google NotebookLM Safeguards Your Information

This is quite significant: Google has made it clear that your NotebookLM inputs, attachments, and chat history are NOT utilised in the training of their global models.

When you set up a new NotebookLM, you effectively create a private and secure space where the AI only knows what you share with it, remaining within that secure environment. This is a vital feature for those dealing with sensitive project plans or technical documentation—it acts as a reliable assistant that doesn’t disclose information.

Mastering the Interface: Sources, Chat, & Studio

To make the most of NotebookLM, it’s essential to grasp the three distinct areas of the user interface. This is more than just a chatbot; it functions as a synthesis engine, delivering structured outputs effortlessly. NotebookLM’s three panels (left = sources, centre = chat, right = studio/output) work as follows:

1. SOURCES: Establishing the Foundations

The left panel is dedicated to Sources. Here’s where you instruct the AI on what information it should consider:

• What to upload: You can easily drag and drop PDFs, Google Docs, URLs, and even YouTube links.
  • Pro Tip: Avoid simply linking to a homepage, as it will only review the text on that page, excluding additional articles or content. If you need specific pages, use the Bulk URL tool and consider employing the Deep Research feature for comprehensive analysis.
• Why it matters: Every response the AI generates will be directly associated with these files, complete with clickable citations. If the required information isn’t in your designated sources, NotebookLM won’t fabricate responses.

2. CHAT: The Reasoning Layer

The centre section serves as the Chat interface. This is your space to interact with the data and ask questions:

• Avoid vague inquiries: Instead of asking “Tell me about solar power,” pose more specific questions, like “Provide a brief overview of the five key constraints for our solar project.”
• Synthesis: You can request comparisons between Source A and Source B; this is particularly beneficial for students uncovering discrepancies in research papers.
  • Some users have found NotebookLM invaluable for in-depth medical analyses, uploading various doctors’ reports (in PDF or TXT format) along with links to videos and articles on relevant issues, and then asking the AI to explain any contradictions.

3. STUDIO OUTPUTS: Transforming Data into Results

The right panel is reserved for Studio Outputs, where you can convert raw data into valuable outcomes:

• Audio Overview: This feature provides a podcast-style discussion of your notes, perfect for reviewing material during your commute.
• Data Tables: This is a personal favourite—by taking a messy transcript, you can instruct the Studio to “Create a Data Table of all action items and deadlines.”
• Video Overviews: You can even create slides and video summaries using diverse visual styles such as Whiteboard or Watercolor.

Understanding NotebookLM’s Limitations

Here are the primary limitations you can expect from Google’s NotebookLM:

Ethical Considerations for Students & Staff

As mentioned in the accompanying video, in the corporate realm, businesses often prioritise high-quality outcomes over how those results are achieved, provided they adhere to standards and avoid data breaches. In contrast, educational institutions such as high schools, universities, and colleges typically adopt a stricter approach.

How To Get Caught Claiming AI as Your Own Work

We have an in-depth article on this topic, but it’s important to note that Google AI offerings like Gemini and NotebookLM contain an invisible watermark called SynthID:

1. Yes, this is a mathematical algorithm woven into the text and graphics.
2. No, it cannot simply be cropped out.

Conclusion

While many AI tools have been hastily brought to market, NotebookLM stands out as it prioritises the privacy that both companies and educational institutions demand. It is an assistant that genuinely works with your data, without disseminating it globally.

If you’re a student or a professional frustrated with unreliable chatbots creating fictitious facts and data, consider trying Google NotebookLM for a change.