AI Podcast Hosts

Google recently announced a new interesting AI tool that could redefine podcasts as we know them. Introducing NotebookLM , a platform that takes files, links, or even text notes and uses AI to transform them into fully produced, 10-minute podcast episodes.

Curious, I decided to give it a try. In just a few clicks, I had a three-minute podcast on this very topic, complete with two AI-generated hosts, one male and one female, who chatted, cracked jokes, and genuinely “sounded” like real podcasters.

Check out the full video HERE !

How It Works:

It’s surprisingly simple. You upload any text format—whether it’s a dense report, a lengthy article, notes from a meeting, or even a website URL—into NotebookLM.

The AI scans the content, extracts the key points, and generates a casual, engaging conversation around it.

My AI hosts interacted as if they were seasoned co-hosts, asking each other questions and adding humor. The voices were incredibly lifelike, giving the episode a polished, professional feel. It’s impressive but also a bit unsettling.

Could we see an influx of AI-only podcasts? How much of the conversation is original versus pulled directly from data? And what exactly is the AI drawing from to make the hosts sound relatable, witty, and knowledgeable?

When it comes to AI, the topic of data, copyright infringement, and what it's being trained on have always been at the forefront of my conversations. Yeah, it’s great; no, I don’t think it will replace humans, but right now, what I believe AI is doing is stealing from us. Stealing our likeliness, our personality traits, our vocabulary, quirks, photography, how we speak, and the art we create.

Every output AI makes today is based on human input, on data collected through algorithms, reports, books, and writing.

This past week, The Atlantic brought up the very topic. Now, AI has a hold on your transcripts. From Google Meet to Zoom, AI chat transcripts are everywhere, capturing sensitive details from business calls, personal conversations, and meetings. This data isn’t just sitting idle—it’s fuel for AI to learn from our interactions.

AI companies store these conversations, shaping how these tools work and raising privacy concerns. With the potential for ad targeting and inevitable glitches, it’s worth considering what we share in AI-driven spaces. These companies never ask for our consent or set clear guidelines for using or monetizing our data.

AI podcasts also make me worry more than I already was about the spread of misinformation. Now podcasts are in the mix, and without a clear visual in place, how will you be able to tell the difference?