It’s no secret that generative AI has been an amazing weapon for spammers, but the scale of some of these operations may make you wince.
A new piece from The Telegraph explores the bizarre new world of AI-generated podcasts — and specifically, a single podcast network called “Quiet Please” that has dreams of online domination. It boasts nearly one hundred separate podcast series, with over 10 million downloads since September 2023.
And it’s just getting started. The company behind “Quiet Please,” Inception Point AI, aims to churn out 5,000 podcasts with more than 3,000 episodes per week, according to The Hollywood Reporter. This would be an extraordinarily expensive venture if the podcasts were human hosted. But using AI, it purportedly costs just $1 per episode, each taking no more than an hour to make. If just 20 people listen to each episode, it’s enough to turn a profit.
Obviously, the podcasts aren’t any good. The network covers a wide breadth of topics, ranging from AI startups, to celebrity biographies, to politics. Many are focused on a single topic, with unintentionally hilarious one-word titles like “Anime,” “Animal,” “Bunkers,” and “Tsunami.”
One spotlighted by The Telegraph is simply called “Lawn,” which is about “the secret world of lawns.” It’s hosted by an AI-generated English gentleman with the ridiculous name of “Nigel Thistledown,” who waxes panegyrical about lawns while telling you little actual information about them. Besides shoveling you ads, it seems mostly a vehicle for the AI to try invent new figures of speeches for the word “lawn.”
“They have attempted to give their presenters personalities, but that seems to mainly boil down to a string of hoary old metaphors and clichés,” writes Chris Bennion for The Telegraph. “Every ‘host,’ whatever accent or tone they are using, speaks in the same monotonous rhythm throughout.”
Even worse, sometimes the AI podcasts are interrupted by glitches so bad that even the most checked out listener might sit up. We return to “Lawn,” per The Telegraph:
“At one point, Thistledown described using technology in your garden as ‘like having a particularly observant but-lutter who anticipates every need,’” writes Bennion. “Presumably to illustrate the sort of help this ‘but-lutter’ gives, a female voice, a posh-sounding, elderly lady, pipes in with, ‘I would never have found my pome!’”
“Thistledown never references this woman and she is never heard from again,” he reports. “We are not told what a pome is.”
In another podcast on the Metaverse, called “Metaverse,” the AI host Leo Finch suffers a verbal breakdown that if produced by a human might be interpreted as a symptom of a stroke.
“That. That sees some. You’re. That sees some. Your. That sees some. That’s not pie conservative point,” intones the AI Finch. “McKinn’s pie. 101’s pie, estimate to make McKinn’s be a challenging. To a two and their critical skills, to home early age.”
It goes to show that nothing’s safe from the rising tide of AI slop. YouTube is overrun with AI generated music playlists, “sleepy” informational videos, and “boring history” pablum. Spotify is a breeding ground for entirely AI-faked musical artists. And let’s be real: since most of us are just zoning out to podcasts anyways, it’s the perfect medium to inject AI into. Unless platforms starting crack down hard on what is clearly just spam with an LLM-veneer, expect to hear more about slop operations like these.
More on AI: President Trump Posts AI Video of Himself Spraying Explosive Diarrhea on Peaceful Protestors
The post Company Churning Out AI Podcasts Filled With Bizarre Glitches They Didn’t Even Catch appeared first on Futurism.
