Over the past three years, I have had the opportunity to write several times about generative artificial intelligence and how it may translate into the meetings and events industry. Starting with an article on the launch of ChatGPT, through pieces on creating images, video and music using prompts in various other tools. I have also told you how Snapsight by Gevme is being used and how you can use Fireflies AI during conversations at trade fairs and send a follow-up within 10 minutes. Many people still associate AI primarily with ChatGPT – although this is changing – and it remains the most commonly used Gen AI tool at work.
NotebookLM by Google
Today I would like to tell you about something completely different, which I have been using for quite some time, and where recent updates have made it even more useful and attractive.
I am talking about NotebookLM. This is a free tool made available to us by Google. One of its advantages, in contrast to other large language models (LLMs), is that when generating content, it relies exclusively on the materials we provide. This means a lower risk of hallucinations, because it is based on our input content. In the so-called “notebook” in the basic version, you can add up to 50 different documents in the form of PDFs, texts, audio files, Google Drive documents or links to YouTube videos.
Based on these sources, within a given notebook, NotebookLM can create summaries, outlines, articles, analyses and compare data. This is real support when working, for example, with public reports spanning many years, allowing you to spot trends or highlight changes. You can ask follow-up questions and extract information that is useful to you.
Importantly, specific data or statements are accompanied by links to the relevant passages in your source materials. Even in text form alone, this significantly shortens the time needed for laborious analysis of multiple documents.
Podcast
What impressed me already a year ago was the ability to create a podcast in NotebookLM. At that time, this was only possible in English; recently, we can also do it in Polish. Imagine – or better yet, test it yourself – a conversation between two people on a topic you have chosen and for which you have provided the source materials. NotebookLM offers four podcast creation options, differentiated in terms of style and purpose.
The first is a detailed discussion: an animated conversation between two presenters who analyse and connect threads drawn from the selected sources, creating a coherent, dynamic narrative.
A short podcast focuses on presenting the key ideas concisely, offering a quick insight into the most important content.
A critical analysis provides a professional evaluation of the materials, enriched with constructive comments and reflections that help to deepen understanding of the subject. The final option, a debate, is a carefully structured exchange of arguments between two presenters, showing different perspectives and interpretations related to the issues under discussion.
My particular favourite version of the podcast is the debate, thanks to which dry text comes to life as a dynamic conversation of arguments and counter-arguments.
The NotebookLM voices sound natural and convincing; sometimes there are still slightly awkward slips or statements and connections between facts which would probably not occur in a normal conversation, but it has to be said that the overall level of the discussion is really good.
In each of the four options, you can also specify in the settings what the AI speakers should focus on in particular. In the English-language version, there is one more excellent option: you can join in as a third person, asking questions and steering the conversation in directions that are more interesting for you.
Audio formats are not the only possibility, because from the same sources you can generate mind maps, video summaries, reports, flashcards and even quizzes. The quiz function proved particularly useful in the “Glossary of Basic MICE Industry Terms and Acronyms” prepared by the Poland Convention Bureau, which allowed users to test their knowledge.
Safety and privacy issues
Of course – I probably do not even need to say this – it is worth thinking carefully about which documents we feed into NotebookLM, both in a private and professional context. The tool can be extremely helpful, but responsible use requires an awareness that the materials we upload may contain sensitive or confidential information, even if Google declares that, for private accounts, it does not use data to train NotebookLM.
Therefore, before adding any file, it is worth making sure that it does not contain personal data, trade secrets or other content that we would not want to disclose.
From my perspective, I use only publicly available materials – reports, studies or analyses that anyone can find online. This way I can be confident that I am not breaching information security rules while at the same time fully leveraging the potential of the tool.
NotebookLM`s multimodality
The greatest advantage of NotebookLM is its multimodality – the ability to work with different types of content: text, audio, images and transcripts. This opens up completely new possibilities for deepening knowledge, as it allows you to combine data from various sources in a coherent, contextual way. Such a multi-dimensional perspective significantly enhances understanding of a topic and inspires new conclusions.
In addition, those who want to broaden their horizons can find recommended public notebooks from other users at the top of the NotebookLM interface. These are often projects or studies prepared by well-known media outlets or online creators, which can serve as inspiration for your own analyses and research. Another advantage? You can also download NotebookLM as an app on your smartphone.
Ultimately, using NotebookLM comes down to maintaining a balance between its technological capabilities and our awareness of what data we feed into the system. If we achieve this balance, the tool becomes genuinely valuable support in both work and learning.
Read also:
