Giacomo Miceli / Ars Technica
This week, an Italian artist and programmer named Giacomo Miceli debuted on the website The Infinite Conversation, an AI-powered non-stop chat between artificial versions of German director Werner Herzog and Slovenian philosopher Slavoj Žižek, complete with realistic voices.
When visiting the site – which is unrelated to either person – you will see AI-generated charcoal portraits of the two men in profile. In between, a transcript of AI-generated text is highlighted in yellow while AI-generated voices simulating those of Herzog or Žižek read through it. The conversation goes back and forth between them, complete with clear accents, and you can skip between each segment by clicking the arrows below the portraits.
The creator positions the site as a social commentary on audio deepfakes and emerging technologies that could undermine trust in the media in the near future. “This project aims to raise awareness about the ease of using tools to synthesize a real voice,” Miceli writes on the site. “Right now any motivated fool can do this with a laptop in his bedroom.”
Giacomo Micelic
Herzog and Žižek seem particularly ripe targets for AI imitation, as listeners may be inclined to believe that the philosophical director and philosopher can say profound things that are difficult to understand. As a result, when the big GPT-3-style language model behind The Infinite Conversation spits out philosophical nonsense, it almost sounds like the real thing. Here’s an example of something the faux Herzog said on the site:
In a sense, Freud also has something to do with literature.
After all, he was a writer.
Yes, he was a scientist, he wanted it
being a scientist, but he was also a writer
who wrote these strange stories.
There is something that seems to be at odds in Freud.
On the one hand, he had such a
strong anthropological vision, which I find very attractive
on the other hand, he was limited in his understanding of cultural history.
He was very antiquarian, for him was antiquity
the most important era because it clearly revealed motivations
while the Middle Ages were just filthy.
Only towards the end of his life did
he sees something good in the Middle Ages.
The dialogue apparently goes on forever. “When you open this website, you are taken to a random point in the dialogue,” Miceli writes. “A new segment of the conversation is added every day. New segments can be generated faster than it takes to listen to them. In theory, this conversation can continue until the end of time.”
Miceli reportedly created the site using “open source tools that are available to everyone,” and declined to provide technical details, though writing on Hacker News that he could produce an explanatory article within a week. “The script generation itself is done using a popular language model tailored to interviews and content written by each of the two speakers,” he writes in the site’s FAQ.
On Ars, we’ve talked about technology before that can manipulate your voice using AI or even allow someone to audibly impersonate someone else. And in October, we saw a podcast using similar speech synthesis technology to enable a fake interview between Steve Jobs and Joe Rogan.
Few will doubt that we are at the forefront of a new era in synthetic media, but if the machines speak, will they make sense? In The Infinite Conversation, that’s not quite the case yet. “Everything you hear is completely machine-generated,” Miceli writes. “The opinions and beliefs expressed do not represent anyone. They are the hallucinations of a slice of silicon.”
