Why is the recent AI victory over a human Go player so significant?

Back in 1997 when a computer defeated a human in the game of chess for the first time ever the whole world was shocked and amazed at the same time. The fear of computers taking over our lives inspired many science fiction writers to come up with stories about the apocalypse and futuristic dystopias where machines rule over all humans. Little did we know that less than a decade later an advanced Artificial Intelligence, which is capable of teaching itself, completely destroys a human in the game of Go.

DeepMind defeats human in game of go

What is Go?

Go is a centuries old territorial game with the goal of controlling more of the board than your opponent. It involves black and white marbles and a table laid out in a perfect 19×19 grid and unlike chess, it has a lot more potential for analytical thinking and intuition. Because of this Go has been thought to be playable only using a human brain since it cannot be brute forced like chess.

What is Google DeepMind?

Google DeepMind, as Techigy explains, is an artificial intelligence software which is capable of learning and more importantly teaching itself to become better and better at a certain task. For example it used to play Go against itself for a long time before the developers felt confident to initiate a match against a professional human Go player.

Unlike the computer that defeated a chess player in 1997, Google DeepMind doesn’t use a brute force approach by going through every possible state of the game before each move. Instead it reduces the amount of moves from that original list and only considers moves that “make sense” in that particular moment. Sort of like a human’s intuition works.

AI victory is a huge milestone

In march 2016 DeepMind defeated Lee Se-Dol, the world’s top Go player, with a score of 4 to 1 in a match that lasted three and a half hours. It’s significant because nobody thought the computer would actually win. After the match Lee himself said he was extremely surprised and didn’t really expect to lose saying “I didn’t think AlphaGo would play the game in such a perfect manner“.

By winning the match DeepMind effectively demonstrated that Artificial Intelligence is more than capable of doing things better than any human would simply because it can learn more efficiently. When you think about it, the average human can play one or two matches a day before he or she grows tired and starts to make errors in judgement, but a computer system can play through hundreds of variations every day and therefore learn more quickly than a human.

The future of Artificial Intelligence

In the near future there will be artificial intelligence systems that can learn and master any skill a human can and because a computer doesn’t get tired and doesn’t require sleep, it can learn everything faster and more efficiently than a human would. On top of that it doesn’t simply “forget” something like we do which means that machines will become better than us in virtually every aspect of our lives.