Google’s DeepMind Program Wins Against A World Champion At The Toughest Game Yet

Artificial Intelligence programs have shown that they can hold their own in games of chess and checkers, but never in Go, an ancient Chinese board game. Before Google’s DeepMind program, AlphaGo, no other AI succeeded in showing more intuition and evaluation skills than a human being. Alpha Go registered a first win last year, against European champion Fan Hui, but last night it truly shone in a match with legendary player Lee  Se-dol.

The first game of Go in Seoul, South Korea saw Lee Se-dol abandoning the game 30 minutes before the official end, when he realised there were no more moves against Google’s intelligent machine. Lee was competing for $1 million prize, so he was more focused than ever, but still the AI met his strategy with unexpected, “deadly” moves. “I was very surprised,” said Lee after the match. “I didn’t expect to lose. [But] I didn’t think AlphaGo would play the game in such a perfect manner.”

For Google, last night’s win was a huge accomplishment and a direct evaluation of their deep neural networks and machine learning system. For Se-dol, it was the first match in a series where he expects to eventually reestablish his reputation as one of the best players of Go in the world.

Next matches will take place tonight, Saturday, Sunday and Tuesday. Until then, we’ll let you watch the confruntation in the video above.