In summary
- Google has launched “NotebookLM”, an AI tool that turns notes and documents into realistic podcasts.
- NotebookLM’s “Audio Summary” update transforms content into lively discussions between two AI hosts.
- AI hosts summarize content and make connections between topics, offering a new way to absorb information.
Listen to this. If we didn’t tell you it was generated (in just a few minutes!), we’d doubt you’d know it was entirely AI-powered.
This is getting wild.
The latest entry in the race to make human creativity obsolete is Google’s AI-powered “NotebookLM,” which offers a feature no one knew they wanted: It can turn your notes and documents into a podcast so realistic that the unsuspecting might mistake it for “Morning Edition.”
The “Audio Overview” update, announced on September 11, transforms uploaded content into intelligently modulated, animated discussions between two AI hosts.
AI hosts dig deeper, summarizing content and drawing connections between topics, giving users a fresh way to absorb information.
That means that, for now at least, an aspiring media mogul could, for example, paste stories about Tuesday’s pager attack in Lebanon from, say, the New York Times, Washington Post, The Guardian, Al Jazeera, Haaretz, CNN and BBC, and within minutes create a 10-minute podcast that rivals any other content out there.
Since this is considered interpretive rather than outright content theft, one could even create a daily podcast from the world’s best sources and sell advertising against it. Here’s what that would sound like.
But podcasts are just the tip of the iceberg. NotebookLM has become a powerhouse for processing and contextualizing large amounts of data. It now handles more than 50 different sources simultaneously, synthesizing information to yield personalized insights. This puts it far ahead of competitors that rely on basic keyword searches.
The update leverages Google’s Gemini 1.5 model, which processes over 1 million tokens. This means richer, more complete contextual insights. For researchers, students, and professionals overwhelmed by data, NotebookLM offers a lifeline, generating study guides, briefing documents, and even inspiring new ideas.
Since its debut in 2023, NotebookLM has gone global and expanded its toolset. It now supports Google Slides and URLs, demonstrating enhanced multimodal capabilities. NotebookLM’s ability to “ground” its AI in user-specific documents seems different than what other chatbots and similar AI tools do by relying on keywords and local searches for their retrieval tasks.
The Audio Summary feature, while innovative, is not without limitations. It is currently only available in English and can introduce inaccuracies. We tested the model with Spanish sources and it produced a coherent podcast… in English. However, in our tests, the model did not generate factual misinformation, it did introduce some contextual elements that could be picked up by wary listeners.
How to make your own podcasts
Generating your own podcasts with the tool is quite easy.
Simply go to the official NotebookLM website and click the “Try NotebookLM” button on the top right corner.
Once inside, you can play around with some pre-loaded examples, interact with them, listen to those podcasts and get familiar with the tool, or you can generate your own podcasts. To do so, click on the tab that says “New Notebook”
Once inside, you’ll find a fairly intuitive user interface. The first thing you need to do is add the sources that NotebookLM will analyze to interact with you via chat or generate the podcast. If you opt for the chat option, the AI will be completely text-based, helping you understand your information and present it better. It won’t affect the podcast format.
You can link your Google Drive account and upload your personal files from there, add a URL from any site you want (it will only understand the text displayed at that specific URL), or directly paste the text you want to analyze.
These options are not mutually exclusive. You can combine them and add up to 50 different fonts.
You can work from the popup window or add sources manually by closing the popup window and clicking the plus button in the left panel. If there is a specific source that you do not want the model to exclude, you can uncheck it.
If you want to see a brief summary of a specific source, simply click on it and you will see an AI-generated paragraph with the most important information, key themes, and the full text of the source below.
If you don’t want to make a podcast from your notes, but are, for example, researching the topic at hand, you can now refer to the notebook you created. Imagine it’s a friend who knows everything there is to know about the information you provided. Go to the little text box at the bottom of the screen and ask your questions.
NotebookLM will give you a very detailed answer. Moreover, its answers will be relevant and will likely surpass any information that a normal AI chatbot would provide. Keep in mind that, like any other AI model, it will sometimes “hallucinate” and provide incorrect information, so if you plan to publish or otherwise disseminate the information, then you will need to verify it.
This one is very powerful, especially for users looking for ways to analyze large amounts of data. For example, you can upload a lot of information about a company and then ask questions to help you decide whether or not it’s wise to invest in it. Or you can upload a bunch of research papers and then ask for relevant information about previous studies on a topic.
But the fun really begins when you ask NotebookLM to generate a podcast about your notes. If you were chatting with the model, simply switch to the “Notebook Guide” button at the bottom right of the screen and click the “Load Conversation” button in the Audio Summary section.
So all you have to do is wait. Within a few minutes, NotebookLM will generate a 10-minute podcast with two AI speakers discussing all the key topics from your sources.
We tested NotebookLM with a real-world use case. Using various sources from the extensive coverage of Donald Trump’s interview on Monday with Rug Radio—Decrypt’s sister company—we fed the information into the system. And the results were pretty good.
In this case, NotebookLM generated fluid content that didn’t veer into mind-blowing territory. It offered solid insights, making connections we hadn’t explicitly made. Internally, our team was impressed with the output, with some unable to distinguish it from a human-written analysis.
However, NotebookLM can’t generate more than 10 minutes of audio, so long-term human-generated content is safe — for now.
Edited by Josh Quittner and Andrew Hayward
Generally Intelligent Newsletter
A weekly AI journey narrated by Gen, a generative AI model.
Crypto Keynote USA
For the Latest Crypto News, Follow ©KeynoteUSA on Twitter Or Google News.