Google NotebookLM’s Raiza Martin and Jason Spielman on the Potential for Source-Grounded AI

This podcast episode of "Training Data" features Raiza Martin and Jason Spielman from Google, who lead the NotebookLM project, an AI-powered research tool. They discuss the origins of NotebookLM, starting as a 20% project within Google's AI Test Kitchen, and its evolution into a viral product known for generating realistic audio overviews or podcasts from uploaded sources. The conversation covers the surprising use cases of NotebookLM, including its adoption by sales teams for knowledge distribution and venture capitalists for analyzing confidential information. Raiza and Jason also share insights into the design choices, challenges, and future directions of NotebookLM, emphasizing the importance of source-grounded AI and co-creation while addressing concerns about AI potentially replacing human roles in podcasting and other professions.