arrow-down arrow-to-left arrow-to-right arrow-up bc-left check delete download facebook google-plus home map menu print search smiling three-lines top-left twitter youtube
August 2017

Why Machines Are Speaking in a Language We Can’t Understand

The rise of AI languages

Researchers at Facebook recently noticed a strange thing happening between artificial intelligence machines: computers are speaking to each other in their own language. This leads us to the question, “what is a language?”

In its simplest form, a language is a system of communication between one subject and another, whether it be a person, a community, or a country. But as machine intelligence increases in complexity, it’s only understandable that the way machines communicate with one another should change as well.

[Facebook] has no way of truly understanding any divergent computer language. “It’s important to remember, there aren’t bilingual speakers of AI and human languages,” says Batra [visiting research scientist from Georgia Tech at Facebook AI Research (FAIR)]. We already don’t generally understand how complex AIs think because we can’t really see inside their thought process. Adding AI-to-AI conversations to this scenario would only make that problem worse.

Although we build machines capable of intelligence, we still don’t understand exactly how they work. And because machine thinking is different from human thinking, the way they approach communication is very different from the way we approach languages.

In this case, Facebook decided to shut down the conversation, due to the fact that the human researchers could not understand what the computers were saying, much in the same way a dolphin and a human might have differences in communication. We are all simply “wired” differently.  

Even when we look solely at the English language, we can see how much it has evolved over the years. Internet lolspeak would be incomprehensible to someone in the 1980s the way ye olde English might raise eyebrows with anyone on the street today. As many great writers break grammar rules after mastering them, languages are ever fluid and prone to change. It reminds us also that there are no “rules” to languages, only standardised practices and publication style preferences set in place so we understand one another better. In other words, languages are never fixed.

Machines can converse with any baseline building blocks they’re offered. That might start with human vocabulary, as with Facebook’s negotiation bots. Or it could start with numbers, or binary codes. But as machines develop meanings, these symbols become “tokens”–they’re imbued with rich meanings. As Dauphin points out, machines might not think as you or I do, but tokens allow them to exchange incredibly complex thoughts through the simplest of symbols.

What does this mean for the future of languages? The definition of the term will remain unchanged, though our perception of it will continue to evolve. As machines become smarter, they will create new methods of communication more akin to our tongues, utilising more words alongside the binary.

“The reason why humans have this idea of decomposition, breaking ideas into simpler concepts, it’s because we have a limit to cognition.” Computers don’t need to simplify concepts. They have the raw horsepower to process them.

Much as how two speakers living in the same country might have a more nuanced understanding of one another because of the shared culture, machines can converse with one another with complex concepts we can’t understand. As we progress, we’ll need to find a way to understand exactly what they’re saying.

What are your thoughts about the future of languages? Share your thoughts on our Facebook Page, and be sure to “like” TELC English for more articles on language learning.

Picture: (c) Fotolia, djama