Google’s virtual podcasts are ready to speak your language

Call us a broken record by now, but the Audio Overviews Google initially cooked up for NotebookLM may just be the coolest thing we’ve seen from the company’s AI efforts over the past couple years. A little over a month ago, Google helped introduce these same podcast-like audio summaries to an even larger audience by baking them in to Gemini itself. It’s already incredibly impressive how the system is able to break topics down and communicate them in an easy-to-grasp manner, and now Google’s got a big expansion to share that’s only going to make Audio Overviews even more accessible.

Being able to parse documents across a swath of languages is impressive enough, but Google’s doing us one better by giving Audio Overviews the ability to generate spoken output in more than 50 languages. Sometimes we’re used to seeing Google get a little piecemeal with adding new languages to systems like this, so it’s great to witness the company instead going all-in with a concerted effort like this — while there will always be edge cases, this should be sufficient to address the vast majority of user needs.

Beyond the sheer scope of how many languages that includes, we’re impressed by the way Google talks about NotebookLM’s ability to seamlessly shift between them. The tool is getting a new setting explicitly for the output language, and if you’re trying to prepare materials for a group with varying language requirements, you can easily toggle that option to whatever you want.

If you have yet to give Audio Overviews a spin, this is as good a time as any. Just fire up NotebookLM, feed it some source documents (or let it search for its own), and let Audio Overviews start weaving its podcast magic for you — you can even steer the conversation in the direction you want.