Google announced in September 2024 the addition of an AI-generated podcast feature to its note-taking app, NotebookLM. The feature, called Audio Overview, converts a user’s research materials into a podcast-style discussion hosted by two AI voices.
Powered by Google’s Gemini AI model, Audio Overview generates audio summaries in which the virtual hosts summarize source material, draw connections between topics, and engage in back-and-forth conversation. The feature builds on text-based summaries that NotebookLM, launched the previous year, had already offered users.
In an internal test, the AI hosts produced a discussion on the invention of the lightbulb, touching on Thomas Edison and the collaborative nature of the invention. The hosts produced remarks such as “in the end, it’s actually a story about teamwork, making the dream work.” The test also revealed some quirks: the hosts spelled out words like “P-L-U-S” and referred to platinum as “bling bling metal.”
Google describes Audio Overview as a reflection of a user’s notes rather than a comprehensive or objective account of any given topic. The company cautions that the AI is not always accurate, and users are advised against relying on it for critical information. The feature is currently available only in English, and generating a podcast can be time-consuming.
To access Audio Overview, users can open a notebook in NotebookLM, select the Notebook guide in the bottom-right corner of the screen, and choose the “Audio Overview” option.
The addition of audio output to a note-taking tool may suggest a broader shift toward more varied content formats for research and study, though the feature remains experimental at this stage.
Source: Tech