Google’s new AI tool, NotebookLM, has the tech world buzzing for good reason. Users of the platform can feed it with any sort of raw data in the form of a text document or PDF – from a simple short stories to staggering reports on complicated subjects – and NotebookLM will turn it into a podcast.
The results? Well, they range from impressive to exciting to frightening.
The “podcast”, hosted by the same male and female voices every time, has the flow and cadence of your typical NPR explainer, with “uhms”, “aahs” and other human-sounding crutch words organically (term used loosely) baked in. After a few “episodes”, it’s easy to pick up how NotebookLM structures the podcast; there’s a pattern to how data is broken down and delivered, and every so often you’ll hear a glitch in someone’s voice, or the “hosts” will repeat entire explanations.
We decided to experiment with NotebookLM by showing it a recent article we posted featuring Kerry Healy, chief commercial officer, Middle East, Africa, Turkey & Asia Pacific for Premium, Midscale & Economy, Accor, and how she’s tackling rate integrity in the Middle East. Prompted by a human, generated by AI, edited by a human.
Have a listen and ask yourself – if you didn’t know this was generated by AI, would you have guessed anyway?
TL;DR – This is not a “real” podcast in the traditional sense of the word. Google has a new Artificial Intelligence tool called Notebook LM that allows users to feed it written data, which it will then turn into a podcast. The voices you hear are completely generated by Notebook’s language models.
