I listened to 200 Notebook LM podcasts so you don't have to
Have you ever wondered what it's like to listen to 200 AI-generated podcasts?
“No. That’s ridiculous. Why would anyone do that?” I hear you say.
I’m flaky on the “why” myself, but the thing is, I did it and those are hours of my life I will never get back.
I listened to 200 fake podcasts generated with NotebookLM to see for myself whether the podcasting industry is about to come crumbling down.
Although I went slightly insane throughout this experience, I feel it was well worth it. I definitely never want to hear the phrase “deep dive” again, but other than that I’ve ventured out of this mess mostly unscathed.
The Genesis
I wrote an article recently about NotebookLM as I had tried it out for the first time. I indulged my inner geek, eager to hear two virtual personas chat about an article that I’d written. Initially, it was a delightful experience—like listening to my words coming back to me through a charming, albeit synthetic, banter.
It all seemed quite harmless and quite a lot of fun to try.
However, over the next couple of weeks, my initial excitement quickly turned into terror when I realised the volume of AI podcasts mushrooming across the podcasting sphere. We're talking 1,466 fake podcasts at last count.
And how did I find all of these fake podcasts?
Enter Listen Notes: The AI Detector
Listen Notes, a trusted podcast search engine that has been around for years was obviously inundated with all of these fake podcasts over a very short period of time.
They weren’t pleased and reached out to NotebookLM to see whether it was possible to provide an AI podcast detector. This would have made it much easier to keep these podcasts off the Listen Notes search platform, but NotebookLM didn’t come to the party.
Thankfully ListenNotes set to work and created a detector themselves which is quite impressive and subsequently how I ended up seeing this list of 1466 podcasts (and now as I’m writing this it’s up to 1781).
And then I decided to listen to 200 of them…
Because that’s a good use of my time clearly.
Here's what stood out:
Deep Dive Overload: Somewhere along the way, AI adopted the phrase "deep dive" as their mantra. I lost track of how many times I heard it and I never want to hear anyone use this phrase ever again. The repetition was maddening. AI needs a thesaurus!
The Boring Banter: While the banter between AI hosts was impressive, it quickly became formulaic. Common phrases like "Hmm, that's interesting" and "I'm intrigued" quickly lost their sheen. Hearing the same inflections and words repeatedly made the experience feel forced.
Lack of Genuine Connection: Podcasts thrive on the connection between hosts and audiences. AI, lacking real personalities or life stories or indeed bodies, fails to create meaningful bonds. Without authenticity, it’s just a bunch of noise.
Trying To Be Relatable With a laugh, I noted claims like, “Ah yes - that old workplace conundrum” or “Sometimes I just feel like being lazy,” or "I've definitely procrastinated myself." This was the only part of the listening experience that genuinely made me laugh. Hearing AI voices talk about their lack of motivation, productivity and how to work in toxic workplaces was actually hilarious. They don’t have productivity issues! Their whole deal is productivity and they’ve never worked in a toxic workplace.
Some AI podcasts felt like experiments, with episodes often dwindling after a few attempts. Without authentic engagement and memorable branding, they fall flat. But if AI could embrace its artificial nature, perhaps humorously acknowledging its own limitations and lack of relatability, that could be an entertaining evolution worth listening to. The day an AI podcast owns its robotic roots and spins it to comedic gold will be the day I eagerly subscribe.
But for now, I'm signing off this virtual podium and returning to the world of human conversations—the ones that come with authentic giggles, quirks, and genuine human connection.
Transcript:
-
[00:00:00] I just listened to so many AI generated podcasts that I want to crawl into a hole and not return for a very long time. But I did it for you. I did it for you. Well, also me. And I want to go through some things that I've learned after listening to so many podcasts. Let's chat about that. let's go back to the beginning.
[00:00:33] So I did a video in a podcast episode where I tried NotebookLM for the first time and got really giddy and excited about the fact that these two fake people were talking about something that I wrote. And it was really cute and fun and silly to then listen to them bantering about some real things that I knew about.
[00:00:54] I thought it was a pretty cool tool. I still think it's a pretty cool tool, but things are getting out of hand. If we then have a look at notebook lm and the amount of actual fake AI podcast. There are now on the podcast platforms for you to listen to. It's 1, 466 at time of recording. That's how many fake podcasts there are out there using Notebook LM.
[00:01:22] Why do I know that? Because ListenNotes Did something about it. They created this detector that you could actually then use to find out whether a podcast was AI generated. They reached out to Google Notebook LM and they said, Hey dudes, this is pretty bollocks. Could you please make it easy for us to be able to detect AI podcasts so that they're not being flooded on the ListenNotes platform?
[00:01:48] And essentially Google was like, mm, soz. Not doing anything about that. And I mean, I love Google, Google. I have Google up the wazoo as far as how I run my business, but [00:02:00] they're not going to spend the time and energy on that at this point in time. So 1, 466 fake podcasts is a pretty big deal. And I listened to a huge amount of them so that I could understand firstly, whether any of them were any good, whether I should be worried as a podcast producer, an editor and podcast creator, whether there is use in the tool.
[00:02:26] And so I'm going to tell you a little bit about what I found after listening to so many of them. First thing I found, I hate Deep dive. I hate the words deep dive. I've heard deep dive so many times now because every single one of these notebook LM podcasts that I've listened to, they are just deep diving up the wazoo.
[00:02:50] Hey, get ready for a new deep dive. Are you ready for this deep dive? Oh, I've got a dive for you. Get ready to deep dive into the deep dive pool. I just, oh my god, I hate it. They just need some other words. There needs to be other words that aren't deep dive. Some of the podcast names are, in fact, deep dive as well.
[00:03:09] Anyway, so I hate it. The words they're using are not that interesting. They have a certain number of words that they've, I guess, been trained on and that's the ones that they like to use. So yeah, hate that one. There's a lot of banter. Okay. They've obviously worked really hard in creating banter between two hosts, which again, I was really impressed with like, man, they've got interjection.
[00:03:32] They've got laughter. The banter is pretty flippin good. They say make sense. Hmm. That's interesting. I like that. I love that. Ooh, exactly. I'm intrigued. These, these, all, all of them, all of the podcasts say that, say those things all the times. And you're probably sitting there thinking like, well, look, Brianna, I'm not sitting there listening to 250 or so podcasts that are all AI generated.
[00:03:59] I might [00:04:00] just listen to one so it doesn't bother me so much. And I mean, I guess you have a point. All I'm saying is, is that if there's 1, 466 and growing, AI generated podcasts and they're all using the same words and they're all using the same inflection. It's gonna get pretty flooded pretty soon full of content that is just the same delivery.
[00:04:20] At time of recording as well there is only two hosts, there are only two voices to be heard. You can't choose them, you can't customize, you can't use your own voice. I know Notebook LM is actually working on you being allowed to customize and choose other voices. That's coming and I guess that will make it a little bit better so that it sounds less templated.
[00:04:40] It sounds less the same as any other AI generated podcast. But, uh, yeah, right now it's just a lot of the same. The other thing I really don't like about them, and this is kind of why I'm not worried about it taking over podcasting and everyone just going, oh, well, I'll just listen to this AI version of the news instead of the real people delivering the news.
[00:05:06] There is absolutely no relationship between the two hosts. You, you can't really establish that you like either of them. They don't have any actual personalities. They've been given a tone that they need to talk in, but you don't feel like you can get to know them, and they don't have insightful thought.
[00:05:29] They don't come in with little, like, stories from their real life, because they don't have a real life. They just spout information at you in the same format that they do for any other podcast. So as I was listening to them, I started to go, at what point would I just not listen to this podcast anymore?
[00:05:48] Maybe I'd listen to a couple of episodes if I felt like the information they were giving was really good and I needed it to be distilled down. But at a certain point, it's just like, what am I listening to you for? Like, it's [00:06:00] boring. Like, it's Kind of interesting. They've made it engaging, but it's still boring because at the heart of podcasting, there is still storytelling and they're not really telling stories about anything that is actually real, like from their own lives.
[00:06:18] That was my first point where I was like, yeah, this is not something I'm actually worried about. Not at this point in time, because at the heart of it, the medium Exists for people to feel connected to other people. Not every genre, you know, some of them are a lot more Just give me the facts but the bulk of people come to podcast to learn something and Connect with the hosts or with other people in the community that are also listening to it and there's got to be an element that like Is bringing everything together.
[00:06:50] And that's normally the person that is delivering the information like me right now. I hope you're connecting with me because I am a person that's got real emotions and feelings and, you know, I can express myself in my own stories and my own thoughts. So AI, no big L M, it does not do that because it can't, it can't do that.
[00:07:11] In fact, I found it kind of funny. There was quite a few instances where they would go like, everyone's craving a bit of authenticity. And I'm like, the, you are an AI, you're not authentic at all. You're a robot. And that was kind of funny. And then they'd say things like, ah, yes, the classic workplace conundrum.
[00:07:30] They've never been in a workplace. They've never had this conundrum. They can't relate. There's so many situations and things in these podcast episodes where they chime in like, Oh yeah, I've had that problem. No, you haven't. You haven't had this problem because you're not. Real. You're not a real thing. You can't have had this situation.
[00:07:50] I don't believe you. And as soon as I don't believe you, I don't trust you. Quite a few of the podcasts were talking about AI and the rise of AI. And I [00:08:00] found it funny when the podcast would be like, you should use AI. And that's why AI is a really good idea. And they're also AI. It's like mind control through AI, getting you to use more AI.
[00:08:12] We use AI every day in our business here at Bambi Media, but you won't see me creating. AI podcasts anytime soon. The other thing that I heard was I've definitely procrastinated a time or two myself. No you didn't! Your whole deal is you don't procrastinate, is that you're super productive because you're a robot.
[00:08:34] I thought that one was particularly funny because it doesn't make Sense. Ugh. Anyway, you can probably tell I'm a bit batty after listening to so much of this notebook LM content, but I did want to showcase a few that I thought, just, just little things that they said that I was like, what are you even doing?
[00:08:54] Oh my God. They've obviously been trained to be relatable, but they can't be relatable at their core because you know, they're robots. What I would like to see is the AI. Owning that their AI. I don't know how you could work that into the model instead of being like, ah, yes I've had that problem or mmm.
[00:09:17] Yeah flippin workplaces, huh? They suck I know I've been in so many of them like instead of doing that It would be really good if they were like, well, yeah I mean, I can't relate because I'm a robot and I don't exist. It would be funnier if And I think I would listen to more of it if they just owned up to the fact that they weren't real.
[00:09:42] That would make for an interesting podcast. I think I would actually enjoy listening to something like that, where they're kind of making fun of the fact that they are also AI and they can't really. Give you any relatable content because they can't relate to basically anything other than [00:10:00] zeros and ones I don't know whether that's how you actually create AI models using zeros and ones Probably not but you know, you get my gist.
[00:10:08] Okay, I'm thinking of Bender from Futurama. That's as far as my knowledge of robots goes So look, that's all I really wanted to discuss today. I just wanted to kind of shed some light on the actual amount, like the sheer volume of fake podcasts that are now entering the market because of Google's Notebook LM.
[00:10:31] I'm not mad about it. I understand that there's going to be new tools all the time. They're going to be widely adopted. Whether this becomes a flop or not, Remains to be seen, but as far as what I've listened to from the sheer amount that I've gone through, they're not relatable, they're not funny, they're not interesting.
[00:10:53] A lot of these podcasts, in fact, most of them, I would say 99 percent of them, don't have any theme music, don't have any significant branding, don't have any way for you to be remembering them later. They just feel like a bunch of bots. I would love to know if you listen to an AI podcast, why do you listen to it?
[00:11:12] Is it good? And can you send me a link to it? Because I would love to listen to it. Bye! Put it in the comments and let me know why you think it's good. At this point in time, I can see it disrupting. I can see something cool coming out of this experiment, but a bunch of AI fake podcasts is not something that I think is going to be continuing.
[00:11:34] A lot of these, in fact, had a very small amount of episodes, three, five, eight episodes, and they haven't released in over a month and people were just trying it out and then. It's just finished. Again, not something I'm super concerned with. I don't love it. I thought it'd be fun to have a listen to them and give my honest feedback on these.
[00:11:57] And the day that an AI owns up to the [00:12:00] fact that it's AI and gives me a podcast that is funny because it's AI, that will be worth listening to.
BAMBY MEDIA LINKS:
GEAR WE USE HERE AT BAMBY MEDIA:
Aputure Amaran 200d LED Video Light
