Google has unveiled a new feature for its AI-powered tool NotebookLM, allowing users to transform their documents into engaging audio discussions. The feature, called Audio Overview, was announced as part of Google's ongoing efforts to improve information comprehension and accessibility.
The feature works by analysing user-uploaded sources and generating a conversation that summarises the material, makes connections between topics, and presents information in a conversational format. Users can access this feature with a single click and even download the audio for on-the-go listening.
NotebookLM, which expanded globally over the summer, has been enhancing its capabilities using Google's Gemini 1.5 multimodal AI model. Recent additions include support for Google Slides and web URLs, improved fact-checking, and the ability to create study guides and briefing documents instantly.
While the Audio Overview feature shows promise, Google acknowledges its experimental nature and current limitations. The AI hosts currently only speak English, may introduce inaccuracies, and cannot be interrupted. Additionally, generating audio for large notebooks can take several minutes.
Google stress that these generated discussions should not be considered comprehensive or objective views of a topic, but rather reflections of the user-uploaded sources. The company is actively seeking user feedback to refine and improve the Audio Overview feature.