AI Podcasts

Back when I was rereading Steal This Book, I did what I usually did, and searched for other reviews on Youtube.  (When you're in the process of forming your own opinions, it's always useful to check them against what other people think.)

There actually were only a couple reviews of Steal This Book on Youtube, one of which was this one:

"Steal This Book" by Abbie Hoffman Book Report Review (Google Deep Dive LLM)

Initially, I believed it to be real, and even thanked the podcasters in a comment:

I just finished reading "Steal This Book," and I found this video to be useful in reviewing and consolidating the book in my mind.  Thank you.

...but, there was another comment from somebody else which read: "You made this new interesting video from AI again."

Was this podcast AI?  It didn't sound like AI to me.  It sounded like two human beings having a conversation.  AI couldn't mimic human conversational voices like that, could it?  Or could it?
There was nothing in the description of the video which indicated it was created by AI, but there was the title of the video, which said "Google Deep Dive LLM".  Did LLM stand for Large Language Model?

I took another listen to that podcast, and actually, maybe it did sound just a tad bit artificial.  At first blush, it seems like normal banter between the man and the woman.  But then, on second listening, they completed each other's thoughts just a little bit too neatly.  Of course, it could easily have been a pre-scripted podcast.  But, if something is going to be this artificially scripted, then could the whole thing also have been AI?

It probably is AI, isn't it?  We all know that AI slop has taken over Youtube and Facebook.  I hadn't thought that AI slop would come for booktube, but I guess in retrospect I should have seen it coming.  After all, there are views to be had reviewing popular books.  Somebody was going to exploit this sooner or later.

I was reminded of this again today when the Youtube algorithm served up to me this video:

When Privilege Masks Dark Actors: From Chomsky to ex‑Prince Andrew! - Without Fear or Favour Podcast

I was listening to it, and finding it interesting and informative for a good 5 minutes or so before I began to get a strong suspicion that this video as well must be AI generated.  The clues?

1) I think it's the exact same man and woman voices as the previous video.  And also a very similar conversational style

2) Looking into the history of this channel, they only started uploading videos 9 months ago--concurrent with the arrival of advanced AI technology.

3) The man and the woman identify themselves as "Jason Carter" and "Sarah Davis" at the beginning of the video, but a Google search for those names doesn't turn up anything about podcasters.  I can't seem to find images or videos of the man and the woman behind the voices anywhere. 

4) This "podcast" doesn't seem to exist on any other podcast apps.

In the end, I feel like this one was pretty conclusively AI.  Still, it wasn't that easy to catch.  It took me a bit.  And I worry that other people might not realize they are listening to AI.  There are two commenters--here and here--who are arguing with the podcast, not seeming to realize the podcast is AI.  (Although someone is responding to those comments.  One wonders what is human and what is AI.  Is a human running the channel, and only using AI to create the podcast episodes?  Or is it AI all the way down.  Is AI responding to the comments?)

Anyway, I don't know how I feel about this whole thing.  Part of me was like "Well, now that I know this whole thing is AI, what's the point of even finishing the video?"

The other part of me was like, "But it still feels like an interesting conversation. Can it still be interesting even though I know it's AI?"

It seems to me that the interest in a podcast is the feeling that you're listening to other humans.  (The whole parasocial connection thing.)  Can I form a parasocial connection with AI?
And if I know that the opinions are not human opinions, but AI opinions, does that invalidate them?  Do I place more weight in an opinion if I know an actual human thought of it?

But if you're worried about these questions, I also worry that in the future, there are going to be a lot more of these AI produced podcasts, and they are going to be harder to spot.  We could be in for a lot more of this in the future.

Given that this video in question is also an opinion piece, I think this also creates some troubling questions for the future.  How many opinion pieces in the future will be generated by robots?  Will robots be shaping human opinions and consensus in ten years?

At any rate, as everyone has noticed, the flood of AI generated content on Youtube and Facebook make these sites a lot less useful.  I used to waste tons of time on Youtube, but at least that was time when I was watching real humans do real human stuff.  Now I keep clicking on videos, and then realizing that those videos are AI.

Well, on the bright side it makes me waste less time on these sites.  Every time I get upset that I keep finding AI videos on Youtube, I always remind myself, "Wait, what am I doing wasting time on Youtube anyway?  I should be reading my book instead."

And, if you're also getting discouraged by AI content, just a friendly reminder that my little corner of the Internet will always be human.  With all my human flaws.  I may mispronounce words on my videos, I may stutter, I may lose my train of thought, I may take forever to get to my point--but through it all, I'll be giving a very human experience.