Google’s DeepMind AI unit has successfully pitched its AlphaGo program against the world’s best Go players. Following a defeat of the European champion last year, AlphaGo has now beaten legendary Korean Go player, Lee Sedol, in the first of five scheduled matches.
Lee Sedol, who resigned after three and a half hours of tense gameplay, is the first 9-dan Go player to face an AI opponent. Ranking second in international titles behind Lee Chang-ho (the only player to have won all eight international Go tournaments at least once), Lee still has four games in which to redeem himself, and he has some strategies on how to do so.
Huge milestone in AI! AlphaGo beats the world champion Lee Sedol! We made history! What an incredible moment! :-) pic.twitter.com/Un4imciEYS
— Mustafa Suleyman (@mustafasuleymn) March 9, 2016
“I am in shock, I admit that….I think I failed on the opening layout so if I do a better job on the opening aspect I think I will be able to increase my probability of winning,” Lee said. The next match takes place on Thursday in Seoul, followed by games on Saturday, Sunday and next Tuesday.
With a $1 million prize for the winner, this tournament is about more than just bragging rights – although AlphaGo’s initial victory is already a historic moment for AI. Especially so considering Go is widely viewed as the penultimate challenge for AI systems for its far greater complexity and reliance on intuition compared to checkers or chess. As Lee noted, “I was very surprised, I didn’t expect to lose. I didn’t think AlphaGo would play the game in such a perfect manner”.
What do you think of this result? Who do you think will win?