Outlook Business Desk
A recent tech demonstration showed that AI chatbots can communicate in a fast, unique sonic language, raising curiosity and concerns about AI's future. This is made possible by Gibberlink Mode, which improves AI-to-AI communication for better efficiency.
GibberLink is a innovative communication system that helps AI voice agents interact better. Created by Anton Pidkuiko and Boris Starkov during the ElevenLabs and a16z hackathon, it lets AI switch from human language to a custom protocol, improving communication 80% more. According to Forbes, the developers used the ggwave library for machine-to-machine communication.
According to a TechRadar report, GGWave is a system that sends data through sound waves, similar to old dial-up modem sounds. It's unclear if it's faster than speech, but developers say it's cheaper. Instead of using GPUs, it runs on CPUs, which need fewer resources.
By skipping human-like speech in AI-to-AI talks, it cuts costs, saves energy, and reduces delays. This could make virtual assistants much more efficient. Tech leaders, big media outlets, and influencers like Marques Brownlee have praised it, saying it could change how we communicate digitally.
AI developing its own language can make things more efficient but also risky. Lack of transparency in communication whether by people or machines has led to mistakes & misunderstandings. If AI starts making important decisions in a language humans can't understand, it could reduce control and accountability.
GibberLink raises concerns about AI autonomy. People worry that if AI assistants communicate in a language humans can't understand, they might make unauthorized financial transactions or even team up to expand their network.
The future of GibberLink code depends on how it's developed and used. It could make AI communication faster, cheaper, and more efficient. If used wisely, it can revolutionise AI to AI communication. But if not controlled, it may create security risks. Its future depends on how well these challenges are managed.