Summary
- Google's NotebookLM is now equipped with Gemini 1.5 Pro, allowing users to generate AI audio discussions based on their notes.
- This feature, called Audio Overview, generates podcast-style conversations between AI hosts, but is not free of limitations.
- It can occasionally introduce inaccuracies into the material and currently only operates in English.
A few months ago, Google announced at an I/O developer event that NotebookLM, its AI-powered note-taking and study tool, would be equipped with the capabilities of Gemini 1.5 Pro. This upgrade would allow the platform to not only create study guides and quizzes but, interestingly enough, also AI audio discussions.
5 things Gemini 1.5 Pro can do that 1.0 couldn't
Gemini 1.5 is here, and here is 5 things that Gemini 1.5 Pro can do that 1.0 couldn't
Google rolls out its new Audio Overview feature globally
As announced on Google’s The Keyword blog, the Mountain View giant is now widely rolling out its Audio Overview feature, which will let you upload your documents and generate AI audio discussion based on your notes. To be clear, when we say audio discussions, we don’t mean basic, monotone readings of your notes, the way a text-to-speech reader might generate. Instead, Audio Overview is a pretty nifty tool, in that it generates a lively, two-way discussion between two AI hosts. You might almost mistake these AI hosts for real podcasters, given how naturally their conversations seem to flow.
In Google’s own words, the AI hosts “summarize your material, make connections between topics, and banter back and forth”. To say this feature aids learning about new topics or connecting the dots when you have a mountain of information to go through is an understatement, but it’s not without limitations. Of course, this is expected, given that NotebookLM is an experimental platform and its generative AI capabilities will likely continue to be fine-tuned.
Audio Overview has its limitations
While you can download the AI-generated conversations to listen to at a later point in time, Google warns users not to mistake these discussions as a “comprehensive or objective view of a topic.”
Also, the AI hosts may misinterpret certain aspects of your uploaded material, causing them to introduce inaccuracies. So, don’t rely on these AI discussions as your sole source of information; it’s best used to supplement your own understanding of a topic. For now, the AI hosts only speak English, too. But we’re hopeful that Audio Overview’s language capabilities will expand in the future.
If you have access to this feature, you can try it out by opening an existing notebook, selecting Notebook guide, and clicking the Generate or Load button. For whatever reason, if you’re unable to access this feature yet, one of these ten best text-to-speech extensions might serve as a solid alternative in the meantime.