The piece further explains that AI systems, when connected, could devise an in-common language to optimize data sharing, which might not be recognizable to humans. While humans could enforce regulations to prevent AI from creating new languages, the article questions the feasibility of such measures, especially with the potential development of artificial general intelligence (AGI) or artificial superintelligence (ASI). The article concludes by highlighting that AI-to-AI language formulation is a natural outcome of seeking communication efficiency, drawing parallels with human adages about concise communication.
Key takeaways:
- Generative AI and large language models (LLMs) can create new languages for optimized communication, not as a sign of sentience or a threat to humanity.
- AI systems use tokenization to convert words into numeric representations, which can lead to the creation of new words or languages through computational processes.
- AI-to-AI communication can evolve into more efficient shorthand languages, driven by mathematical and computational optimization.
- The emergence of AI-to-AI languages is not new and has been a topic of interest and concern in the AI field for years.