The latest update to Google’s AI-powered research tool, NotebookLM, can turn your deep reading material into a podcast-like audio discussion. Called Audio Overview, the new feature takes information from documents you upload and generates a “deep dive” discussion with two AI hosts. Google says that in addition to summarizing sources, the hosts can also find links between different topics and joke with each other. Based on the examples the company posted in its announcement, the AI hosts sounded human enough to listen, though you can still tell from intonation and odd pronunciation of certain words that the audio was generated by an AI.
Because the feature is still experimental, Google acknowledges that it has limitations. Currently, hosts only speak English and may speak inaccurate information, which means you have to double-check your material to make sure you don’t learn something that’s not true. You also can’t interrupt a host while they’re talking. And it still takes a few minutes for NotebookLM to generate an audio summary for notebooks with large files. In the feature’s announcement post, Google Labs product manager Biao Wang wrote that despite these limitations, the team is “excited to bring audio to NotebookLM” because “we know that some people learn and remember better by listening to conversations.”
In 2023, the company released NotebookLM, a digital assistant-like app that can ask questions about uploaded documents. In June of this year, Google announced that NotebookLM was officially launched with Gemini 1.5 Pro, adding new features and tools and expanding to more than 200 countries and regions.
