Artificial Intelligence has been making the rounds recently in the media, with Elon Musk, Steven Hawking, Bill Gates, and many others sharing their thoughts and concerns about AI, saying it is a potential threat to humanity or will create some other nefarious consequence. Many media outlets have offered additional commentary attempting to reduce potential panic, suggesting concerns are overblown, and that AI will ultimately be a benign and beneficial phenomenon.
Curiously absent from the whole discussion is a very pertinent and critical aspect of Artificial Intelligence, the Language Barrier.
When AI pops up, the side that will struggle is the humans, not the computers. (Regarding the core debate over the existential possibilities of AI in the first place, please see my note below)
In the 2013 movie ‘Her’, with Joaquin Phoenix and the voice of Scarlett Johansson, there is a scene that is amazingly indicative of our potential future with Artificial Intelligence. I won’t give away too much in case you haven’t seen the movie, but in a climactic point, Samantha (Johansson’s character), the Artificial Intelligence, says to Theodore (Phoenix):
it’s like I’m reading a book… and it’s a book I deeply love. But I’m reading it slowly now. So the words are really far apart and the spaces between the words are almost infinite. I can still feel you… and the words of our story… but it’s in this endless space between the words that I’m finding myself now.
Taking out the romantic angle, the AI Samantha is able to operate, think, and communicate so much faster than her beloved human Theodore. Samantha is not human, she is not bound by our limits.
Humans are extremely complex and amazingly powerful computers, which operate in many ways far beyond the current capabilities of the most sophisticated systems in the world. Think about it, look out the window. Immediately, and intuitively, you know what a person is, approximately how old they are, what time of day it is, whether a window is open (and what that means), and so many other elements of the world around you. Computer systems, in many way are no where close to those capabilities. Don’t get me wrong, there are some very sophisticated visual systems that can tell me the kind of shoes I’m wearing, but they aren’t quite ready to understand deeper contextual clues that humans know intuitively.
But that is on the visual side – Let’s shift to communications. Take a moment and as fast as you can, state your full name and your full formal address. It will, at best, take a few seconds. In the time that you shared two rather mundane facts, two computer systems, with run-of-the-mill gigabit ethernet connections, could have shared the entirety of a DVD’s worth of Breaking Bad episodes.
The simple fact is that computers operate and can communicate at a speed and in a way that humans can’t even attempt to match. A great demonstration of this style difference is the annual International Obfuscated C Code Contest (http://ioccc.org). This is a contest where programmers are awarded for code that produces functional programs in a way that appears to make no sense or can not be easily understood by humans. Let’s put that another way: Programs that we cannot easily understand that operate just fine, and generate useful, albeit simple, spreadsheet or calculator programs.
After seeing that, it’s logical to conclude that a truly artificial intelligence will seek to operate in the most efficient way possible. They will not be bound by our long-form, context and inflection heavy communication style. Because of that, humans will need extensive support mechanisms to communicate with the computers, not vice-versa.
In an AI world, we may chat to our neighbor about how their dog dug up our yard again, while an AI-enabled smartphone in our pocket spends it’s idle time optimizing our commute to reduce carbon emissions and give us more time at home without telling us. Who will need to come down to the other’s level?
With this in mind, I think we should all generally relax and not worry about artificial intelligence. If our AI overlords decide to act against humanity or do something nefarious, we will most likely have absolutely no idea what’s going on or even be capable of grasping what’s happening before it’s too late.
Now, about that dog….
Note: I have put aside the core debate over whether true Artificial Intelligence can and will arrive in any sort of reasonable time frame. There are many rational and reasonable arguments for why AI will not appear anytime soon, e.g. algorithms do not equal AI & and many others, and many doubt that AI is even possible. Personally, I believe that Artificial Intelligence will arrive sooner than later, and for the sake of this article, I have assumed that it can and probably will appear.