Google's AI tool turns YouTube videos into podcasts – here's how to try it yourself

As an avid listener of podcasts who also loves a good, in-depth informational YouTube video, I suspect that Google’s NotebookLM latest update can be oddly satisfying for those interested in the best of both worlds.  

NotebookLM, Google’s artificial intelligence (AI)-powered notebook, now enables users to convert lengthy YouTube videos into “concise AI-generated podcasts, saving time and increasing study efficiency.” This latest NotebookLM capability comes on the heels of September update that allows users to add public YouTube URLs and audio files directly into their notebook and generate study guides. 

Also: Google’s AI podcast tool transforms your text into stunningly lifelike audio – for free

While this update is not suited for vlogs or GRWM videos, if you watch debates, conferences, interviews, and other such informational content on YouTube, this could be a handy tool — especially if you enjoy digesting new information while commuting long distances or, say, spending the day at the park. Essentially, you can curate your own podcast episodes and listen on the go, which isn’t always possible with a YouTube video if you’re not connected to WiFi. 

NotebookLM is still in the experimental phase, but the digital notebook has been expanding its source base since its 2023 launch. NotebookLM collects notes — called “sources” — which can be PDFs, Google Docs or Slides, pasted text, audio files, YouTube links, or web URLs.

How to turn YouTube videos into AI podcasts 

First, visit the NotebookLM website or Google Labs and click Try NotebookLM. You’ll need a Google account to sign in and create a new notebook or to access an existing one. 

Also: Google’s NotebookLM can discuss your notes with you now. How to access it (and why you should)

Click on “Link” in the source selection area, choose “YouTube” and paste your desired YouTube video URL. Select “Generate” in the Audio Overview section to create your AI podcast.

Google has cautioned that the Audio Overview feature generating the AI podcast has limitations; for example, it can take several minutes to generate an Audio Overview for larger notebooks, and the results can contain inaccuracies.