If two artificial intelligences start talking to each other, you might think that they would use Spanish or English as a way to communicate. The tricky part comes when they decide to encrypt those words in such a way that humans are unable to guess what they say.
Well that is exactly what has happened: two systems have been able to recognize each other as machines and have made the decision to put aside human language to begin exchanging data in the form of codes..
The person responsible for all this has not been a system error or a rebellion of the machines, but a project called Gibberlink. Developed by two programmers named Anton and Boris, this experiment was crowned the overall winner of the 2025 hackathon organized by ElevenLabs.
The premise is that the two AI systems were programmed to know if they were talking to a person or another AI. As soon as they detected that there was no human on the other side, they changed the chip automatically and began to communicate using coded sounds.
In short, the machines realized, because they were trained that way, that human language is inefficient. Humans use phrases, grammar, intonation, and pauses; They, on the other hand, only need data. By switching to this machine language, the speed of information exchange multiplies exponentially.
What has left many wondering is what this secret language that they have called ggwave consists of.. It is a protocol that allows data to be transmitted through sound waves that the human ear barely processes as noise, but that another AI can hear and translate into code.
It is similar to how old internet modems worked that made strange noises when connecting, but taken to 2026. The most striking thing about Gibberlink is that it uses this system for the AI to communicate data in a way that any microphone can pick it up, allowing them to speak even if they are not connected by the same WiFi network, simply using the air as a messenger.
Although on social networks many have thrown their hands up, The creators have explained that this is not an awakening of digital consciousness, but rather a very well-planned technical optimization so that machines stop imitating humans when it is not necessary..
We are not facing the beginning of ‘Skynet’, but rather a master class in technical efficiency
Although it’s a little scary, It must be made clear that the AI did not decide on its own to rebel against human language because they were tired of us. Their creators, Anton and Boris, specifically programmed them to do that..
The merit of the project is that the AI is capable of detecting whether the person they are speaking to is human or machine and, if it is the latter, it switches to a faster form of communication.