It was just another day at the Facebook office. Bob and Alice were having a conversation.
Bob: “I can can I I everything else.”
Alice: “Balls have zero to me to me to me to me to me to me to me to me to.”
That gibberish might be the most sophisticated conversation on earth.
Let’s set something straight here. Bob and Alice are not human. They are two of the many AI-powered chatbots developed by Facebook, and they are conversing in the “more enhanced English” developed by themselves.
The problem is the language is nothing like the one we use and understand. If the machine is starting to talk behind our back, should we be worried?
Chatbots and Their Need for Faster Learning
The things that we call chatbots have been around since the era of IRC or even earlier. But they recently became popular after Apple announced Siri, the intelligent personal voice assistant. After that everybody was trying to develop their own version of an AI-powered voice assistant, like Amazon’s Alexa, Microsoft’s Cortana, Facebook’s M, Soundhound’s Hound, Viv, Ozlo, X.AI, and many more – which are either already out there or still in development. Mark Zuckerberg even created an AI personal assistant for himself voiced by Morgan Freeman.
The goal is to make our lives easier. The chatbots will assist users from playing music to giving financial advice. And to do that the creators have to train them to be smarter, to provide faster responses, and to better predict users’ needs. And sometimes the scientists might need to let the bots chat with one another.
But apparently, some of them think that the language they use is not enough to convey their message, so they developed their own language and started to have “private conversations.”
The Terminator and the Realm of Sci-Fi
The first time I stumbled into the news, I couldn’t help but think about Skynet and the Terminator stories. For those who have been living under a rock for the past few decades, the story is about a world where intelligent machines have taken over the human world and are trying to terminate all the remaining people.
Of course, that’s pure science fiction. But history has shown us time and time again that much of science fiction does become a reality. Here are a few of the examples that pop up in my head: flying machines, GPS, tablet computers, driverless cars, and smart homes. Heck, even the things that we take for granted today, like the Internet and drawing on a touch screen, were science fiction once.
That’s different, you might say. All the things that I listed above are basically possible. But machines taking over the world and destroying humans?
Well, to put things into perspective, try showing your iPhone to the people from the 60s (if you can get access to one of the soon-to-be-invented time machines). The 60s generation are the ones that landed on the moon. But to them a supercomputer that can fit into their pocket – and play augmented reality games like Pokemon Go – would be an impossible device that belongs to the pages of sci-fi novels.
Soon we might need to start looking for our John Connor.
Elon Musk and the Fear of AI
Here is the man who pioneered digital payment, smart electric cars, super batteries for homes and vehicles, ultra-fast land transportation, and even wants to put men on Mars with his cheap space travel; Elon Musk is no stranger to walking on the edge of technology. He might soon invent his own Iron Man suit and become the real life Tony Stark. But this same man is also stressing that governments need to start regulating AI.
Some might say he’s over paranoid. But he said, “I have exposure to the very cutting edge AI, and I think people should be really concerned about it.” And if you think about it, Elon Musk and his companies would probably profit a lot more from the lack of AI regulations, yet he’s actively encouraging the government to push the emergency stop button, just in case.
Just as with everything else, Artificial Intelligence is developed to make human lives easier. What started as sets of pre-configured reactions to specific actions, AI-powered machines have beaten the best human in the games of Chess and Go and possibly in everything else in the near future.
It’s undeniable that robots do everything better than humans. They’re precise, on time, tireless, and require no vacation nor paycheck. It’s only a matter of time before they think they are better suited for this world than us – if we let them.
Bob and Alice have just shown us that we might have opened the door to that in the future.
Bob, Alice, Mark, and Interconnectivity
Now, before you start building your underground bunker, please note that the scientists at Facebook AI Research (FAIR) have stopped Bob and Alice as soon as they realized these two bots have strayed from plain old English. So, I guess we have some additional time before the rise of the machine.
But not everybody is living on the gloomy and dark side of the world. For example, Facebook’s head honchos thought Elon Musk’s view on artificial intelligence was “pretty irresponsible.” There are so many good things that can come up from the advancement of AI, and there are already many examples where AI improves human lives.
Some others argue that we should let the AIs continue developing their own variations of language. One reasoning is for them to come up with more effective ways to communicate between machines.
Similar to other products of human evolution, English – the base language that we teach the chatbot to speak – is full of irregularities. While the language is fine for humans, it’s imperfect, making it inefficient for machines. Teaching two AI chatbots to use English to converse is like asking top notch programmers to use bad codes. At the end of the day they will come up with the simplest and fastest alternatives.
Letting the chatbots continue in their foray might bring us to a new era where machine language is faster and more efficient, compatible with more types of devices, and able to boost the better overall performance of everything. Imagine everything around us working better, from toasters to refrigerators, traffic lights to smart cars, personal computers to printers to hard drives, and virtually everything else.
The world where machines can think for themselves with minimal human interference do human jobs better, communicate faster with each other with the language only they can understand. Wouldn’t that be a better world?
Or is it downright scary?
And don’t forget to share your opinion in the comments below.