Google’s AlphaGo AI, trained to play the ancient Chinese game of go, beat one of the game’s world champions, Lee Sedol, in a televised match. AlphaGo won the first game, which Google estimates had over 60 million viewers on Chinese television, and went on to win three more out of the five-game match-up.
Computers have dominated chess, with its relatively fewer number of possible moves, since 1996, but go had long been considered too complex for computers to master. Instead of being programmed to play, Google’s AlphaGo was programmed to learn and then taught how to play. After playing through millions of games and learning from its mistakes, it was finally ready to take on a world champion.
Lee’s win over AlphaGo in the fourth match shows the machine was fallible, and after the matches, observers were able to comment on the strengths and weaknesses of various moves AlphaGo made; however, the artificial intelligence often surprised its opponent with moves that “no human would have made” but that nonetheless helped secure its victories.
To read more articles from the Spring 2016 issue, subscribe to our eReader edition.