What happens when the stuff of bizarre thought experiments becomes something that actually exists? That’s a question that’s come up with increasing frequency this century, and while we’re not quite at Torment Nexus levels, one enterprising and tech-savvy guy has figured out how to make an AI version of controversial podcast host Joe Rogan. “What would an AI version of Joe Rogan sound like?” you might ask. That is no longer a rhetorical question. And if you’ve ever wanted to hear a simulated version of Rogan interview a simulated version of Donald Trump, this is now a thing you can do.
Rogan himself seems a little unnerved by the existence of his AI doppelgänger. “This is going to get very slippery, kids,” he tweeted in response to the first episode.
At The Verge, Amrita Khalid spoke with Hugo (who asked to only use his first name), the man behind The Joe Rogan AI Experience, who described the process of bringing this real podcast featuring fake versions of real people to fruition. As Hugo explained to Khalid, the process involved generating dialogue with ChatGPT and then using what Khalid described as “a text-to-speech platform that can clone any voice based on snippets of audio.”
Deepfakes Are on Reality TV Now. What Could Go Wrong?
The creators say this could make the internet more “human.” Critics disagree.Hugo went on to explain that the third episode took “more than three weeks” to produce — in part because of the challenges of getting something that sounds authentic. “If it’s not good enough in my eyes, I will spend a lot of time making it right,” he told The Verge.
Neither Khalid nor Hugo seems to think that The Joe Rogan AI Experience will ever be indistinguishable from the real Joe Rogan Experience podcast, which is encouraging. But as with many things involving AI and deepfakes, it’s unnerving to think that if this is what one person can do in their spare time, what could a more dedicated effort produce?
Thanks for reading InsideHook. Sign up for our daily newsletter and be in the know.