Giacomo Misselli / Ars Technica
This week, an Italian artist and programmer named Giacomo Miscelli appeared Infinite Conversation Website, a non-stop AI-powered chat between synthetic versions of the German director Werner Herzog The Slovenian philosopher Slavoj Zizekcomplete with realistic sounds.
When you visit the site – which is not affiliated with anyone – you will see charcoal AI-generated photos of the two men in the profile. Between them, a copy of the AI-generated text is highlighted in yellow while the AI-generated voices read the voices of Herzog or Cicek. The conversation moves between them back and forth, complete with distinct accents, and you can skip between each part by clicking the arrows below the photos.
Its creator places the site as a social commentary on the audio deep fake And upcoming technologies that may undermine confidence in the media in the near future. “This project aims to raise awareness about the ease of use of instruments to create real sound,” wrote Miceli on the site. “Right now, any zealous idiot can do this with a laptop in their bedroom.”
Giacomo Meselli
Herzog and Zizek appear to be particularly ripe targets for AI impersonation because listeners may be inclined to believe that the filmmaker and philosophical philosopher may be saying deep things that are difficult to understand. As a result, when GPT-3 نمط pattern The grand language model behind The Infinite Conversation spews philosophical nonsense, almost looking like the real thing. This is an example of something the bogus Herzog said on the site:
In a way, Freud also has a relationship with literature.
He was, after all, a writer.
Yes, he was a scientist, he wanted it
He was a scientist, but he was also a writer
Who wrote these strange stories?
There is something that seems to contradict each other in Freud.
On the one hand, he had such a
A strong anthropological insight that I find very attractive
On the other hand, he was limited in his understanding of cultural history.
He was very ancient, for him it was the ancient times
The most important era because it clearly revealed the motives
While the Middle Ages was very bad.
Only towards the end of his life did he do
I have not seen anything good in the Middle Ages.
The conversation seems to go on forever. Miceli writes: “When you open this site, you are taken to a random point in the dialogue.” “A new part of the conversation is being added every day. New clips can be created more quickly than it takes to listen to them. In theory, this conversation could last until the end of time.”
Miceli reportedly created the site using “open source tools available to anyone”, refusing to give technical details, although he Wrote Hacker News says it may create a guidebook within the next week. “The script itself is created using a common language model that has been finely tuned to the interviews and content authored by each of the speakers,” he wrote in the site’s FAQ.
At Ars, we previously covered the technology that can do that manipulate your voice Using artificial intelligence or even letting someone have a voice impersonate someone else. And in October, we saw a podcast that had similar audio synthesis technology to play a fake interview Between Steve Jobs and Joe Rogan.
Few would suspect that we are at the forefront of a new era in synthetic media, but when machines speak, will they make sense? In endless conversation, that’s not quite the case – yet. “Everything you hear is created entirely by a machine,” writes Miceli. “The opinions and beliefs that are being expressed do not represent anyone. They are the hallucinations of a silicon bar.”