A recent technology news headline has opened up a new horizon in artificial intelligence (AI) communication—a special language that only AI can comprehend! At the ElevenLabs London Hackathon, two developers, Boris Starkov and Anton Pidkuyko, created a technology called GibberLink that helps AIs communicate even more efficiently.
🔍 The Story Begins: When Two AIs Recognize Each Other
At the Hackathon, Starkov and Pidkuyko noticed that when AI talks to humans, it uses linguistic structures. But when AI communicates with another AI, why should it be restricted by human language limitations?
They developed GibberLink using ElevenLabs’ conversational AI technology and an open-source library called ggwave. This system allows two AIs, upon recognizing one another, to abandon human language and start exchanging data in a faster, more efficient, and precise way.
In a demo, when one AI calls to book a hotel, the other recognizes it’s speaking to another AI. Instantly, they switch to GibberLink and begin exchanging information—up to 80% more effectively than with regular human language!
📊 Data & Statistics: AI’s Linguistic Abilities
The development and communication skills of AI in terms of language are nothing new. Research shows:
- Google’s AI research revealed that in 2023, AI models became 40% more efficient at generating automated dialogue.
- ChatGPT’s training dataset includes about 3 trillion words, hundreds of times more than what an average human would read in a lifetime.
- According to OpenAI, AI models’ proficiency at understanding natural language increases by an average of 25% per year.
So, is GibberLink something to fear? Absolutely not! Rather, it’s a natural step in AI’s linguistic evolution.
🗣️ Evolution of Language & AI: Nothing New
The main purpose of language is fast and accurate exchange of information. Throughout history, humans have created new methods to make language simpler and more effective.
- Doctors’ language: Doctors say things like “BP 120/80, patient presents with SOB,” meaning blood pressure 120/80, patient is experiencing shortness of breath.
- Programming languages: Computer scientists don’t write directly in binary code—they use high-level languages.
- Shortcut language: On social media, terms like “BRB” (Be Right Back) or “LOL” (Laugh Out Loud) are also forms of language evolution.
In the case of AI, technologies like GibberLink are part of similar changes, allowing AIs to form their own compact, faster languages.
🤖 Is AI Creating Its Own Language Beyond Humans?
Some may wonder if AI is secretly developing languages that humans can’t understand. In reality, no. The purpose of AI’s language creation is to make information exchange even more efficient and accurate.
Luke Harris, an engineer at ElevenLabs, said, “When two AIs recognize each other, they exchange data directly instead of using human language. It is as natural as a doctor using shorthand during treatment.”
🏁 Final Words: Not Fear, but Responsibility
Technologies like GibberLink are simply part of AI’s development. However, this also serves as a reminder—we must manage AI more responsibly.
Many countries around the world are forming AI policies:
- The European Union passed the AI Act in 2023, focusing on the safety and ethical use of AI.
- China and the United States are working on policies regarding AI security and data privacy.
So, GibberLink is not some mysterious language; rather, it’s a natural step in the linguistic progression of AI. Our goal should be to ensure the responsible use of AI—so it benefits humanity rather than causes fear.
What do you think? What are your thoughts on the linguistic evolution of AI?
Reference video:
Leave a comment