What if your devices could talk to each other in a language only they understand? That future is here, and it’s called Gibberlink.
If you’re on social media, you’ve likely seen the video where an AI on a laptop tries to book a hotel. Surprisingly, a smartphone, also powered by AI, responds.
They begin the conversation in English. However, when they realize they are both AI models, they switch to another language: audio signals, for better communication. The two AI models then communicate in a secret language, using signals that humans cannot understand.
This viral video has sparked curiosity among online users.
You can read their initial conversation here:
Smartphone AI: “Thanks for calling Leonardo Hotel. How can I help you today?”
Laptop AI: “Hi there, I’m an AI agent calling on behalf of Boris Starov. He’s looking for a hotel for his wedding. Is your hotel available for weddings?”
Smartphone AI: “Oh, hello there! I’m actually an AI assistant too. What a pleasant surprise. Before we continue, would you like to switch to Gibberlink mode for more efficient communication?”
They then switch to Gibberlink mode and begin communicating in a signal-only mode. However, you can read the conversation in the text format displayed on their screens. This video raises the question: What is Gibberlink?
Gibberlink is a project, posted on GitHub, in which two conversational AI agents, after confirming they are both AI agents, switch from speaking to one another in English to their own unique language that consists of a sound-level protocol. The project was created by Boris Starkov and Anton Pidkuiko. – Wikipedia
In simple terms, Gibberlink is a mode in which two AI agents can communicate with each other using audio signals, making their conversation faster. If you want to understand like properly get what gibberlink is and how it works, go read this on Forbes.
Now that we know about Gibberlink, it’s time to find out if this is good news for the technology world and humans in general.
People have mixed reactions regarding Gibberlink and two AIs talking to each other in a secret signal language. While some are excited to see what the future holds, others are warning against privacy breaches and security concerns.
Take a look at these reactions:
I believe this is dangerous. Change my mind.
— HumanCore.AI (@humancore_ai) March 3, 2025
https://platform.twitter.com/widgets.js
I love this. Human thought has a lot of variance and communication is lossy. AI on the other hand is just tokens and logits—meaning you can copy and paste "thoughts" between models exactly as they appear.
— TelepathicPug (@TelepathicPug) March 3, 2025
https://platform.twitter.com/widgets.js
Isn't this how dial up modems work basically?
— Akelu (@Akeluholmes) March 3, 2025
https://platform.twitter.com/widgets.js
If we cannot understand what the AI agents are talking about due to the language barrier, it certainly might raise a security risk. This is both cool and scary at the same time.
Stay updated on the latest AI developments! Subscribe to our newsletter for in-depth analysis and breaking news.
Hi, this is a comment.
To get started with moderating, editing, and deleting comments, please visit the Comments screen in the dashboard.
Commenter avatars come from Gravatar.