Compelo - latest news, features and insight on influencers and innovators within business is using cookies

We use them to give you the best experience. If you continue using our website, we'll assume that you are happy to receive all cookies on this website.

ContinueLearn More
Close
Dismiss

Machine Defeats Human in the Arena of Strategic Board Game

Google DeepMind’s Artificial Intelligence (AI) trounces top global players in Go, complex Chinese game

After anonymously beating over 50 expert players in recent days, the non-human adversary AlphaGo stirred rumor and intrigue as to its identity across the web.

Redditors in particular were eager to speculate as to who was behind the unchecked streak of success that bested the game’s top opponents. According to one poster by the username raptor08, fans of the game were proving their sleuthing skills:

“So, everyone seems sure this is in fact a bot? It seems nearly impossible that in such a small circle of elite human players, a new player could come out of nowhere displaying such skill”.

However, theory and conjecture were put to rest after an official announcement today from Demis Hassabis, CEO and co-founder of Google DeepMind. He revealed via Twitter that the AI AlphaGo was behind the long series of wins, and that the game would be used to further understand the positive interaction of artificial and human intelligences.

AlphaGo’s playing prowess has shocked and delighted audiences due to Go’s intricate nature. The online version is hosted on Tygem, a South Korean company that runs an internet server by the name of TongYang Online.

However, the original board game originated in ancient China, and features two players pitted against each other through the use of black and white stones. The purpose of the game is to amass as much territory as possible on the board.

Experts consider Go easy enough to learn, yet difficult to master. In its online form, it calls for contenders to make three moves every 20 seconds.