This week was a pretty big news week. With the Iowa caucus coming up, everyone wants to know exactly what is in those 22 Clinton emails deemed too classified or otherwise are too caught up with the Trump scandal (which one exactly is up to you). Flint, MI and the Malheur National Wildlife Refuge in Oregon seem to be the only places in the United States anyone wants to read about. But, after weeding through all of that you think you would have seen at least one measly thing about Google making a technological advancement seen only in Hollywood films up to this point, right?
So, why are you only reading about it now? Why is this not plastered on the front page of every newspaper, every Twitter timeline, and every news source?
Well, first of all, this is not some full blown version of Her or anything. That being said, Google’s artificial intelligent system AlphaGo is nothing to scoff at. The system is designed to beat Go, a game of strategy infinitely more complex than chess. The term “infinite” is not being used lightly here either. The average turn in chess produces about 35 possible moves. Go produces 250. And then another 250 after that. And so on ad infinitum until the end of the game. This is not simply a supercomputer. A supercomputer does not have the storage capacity. There are “more Go positions than atoms in the universe.” In order to play Go, a system has to be able to have intuition and learning abilities. A system has to have intelligence.
Go is the ultimate test for AI. Up until this point, other AI systems have only been able to be matched against amateurs and have proved to be no competition for grandmasters. AlphaGo just beat the European champion 5-0; a feat previously thought to be at least another decade away. Google is confident that their system is going to beat the top Go player in the world, Lee Sedol, this March, proving that AI has made the first huge advancement into human thought.
This all sounds like a pretty big deal, does it not? Well, before you go out and buy "Ex Machina," remember this is just a board game (yes, a very complex game directly linked to human intelligence, but still just a board game) and the Google DeepMind research team has complete control over AlphaGo. The system is also designed to maintain a relatively limited universe. Although it can learn and even think by itself, it can only do this within the universe programmed for it (in this case the universe of Go). However, this could lead to huge advancements in the world of technology. Imagine a translation system that can learn and think in any language; translation error would become a thing of the past; learning languages would become a distant memory. Still worried? Keep your eye on Prof. Nick Bostrom and the Oxford Future of Humanity Institute. These professors are dedicated to asking the hard questions as to where exactly humanity is headed with artificial intelligence and pressuring scientists on the ethical concerns of AI.