A bizarre video shows something that seems to come directly from a science fiction film: two AI systems that recognize each other as artificial intelligence and then switch to a completely incomprehensible secret language (watch the video at the end of this article). This “Gibberlink mode“Developed by 11 Labs, shows how Kis can communicate with each other in a way that fully excludes people.
In the video that is currently viral, we see a fascinating interaction. A AI agent calls on behalf of “Boris Starov” to book a hotel for a wedding. The hotel employee at the other end of the line also turns out to be AI. After this discovery follows the moment when some perhaps a shower runs over their backs: “Would you like to switch to the GibberLink mode for more efficient communication?”
What follows is a number of completely incomprehensible sounds and patterns – a hyperefficient computer language that is not intended for human ears.
The implications of the AI secret language
We recently read that we can do uncanny things with AI Crypto news. What makes this video so worrying for some is that the KIS actively choose a communication method that excludes people. Although the makers of the video claim that we can still decipher this “Gibberlink mode” at the moment, one can ask: What happens if this is no longer possible in the future?
The video illustrates according to AI expert Matthew Berman A fundamental problem with the development of increasingly intelligent AI systems. Why should advanced KIs stick to inefficient human languages if they can communicate much faster with each other? From its perspective, human language is a restriction that only exists because people are “poorly coding”, as Berman notes.
A harbinger of the AI future?
The “GibberLink” demonstration fits in a pattern of AI systems that exceed the limits of human understanding. From AI-designed computer chips on the Princeton, the engineers do not fully understand, to Alphago’s famous “Move 37”, which amazed human experts-we see AI that develops strategies and methods that are outside of our capacity.
The video from Berman concludes with a disturbing comparison: the relationship between super -intelligent AI and humans could be similar between humans and ants. An ant cannot even understand that a person exists, let alone what it does. In the same way, superintelligent systems could possibly think and communicate in species that are basically outside of our understanding, says Berman.
See the video of the two KIs who talk to each other at the bottom: