Google’s ‘Go’ Victory Signifies AI Is Eclipsing Human Capabilities

Google's 'Go' Victory Signifies AI Is Eclipsing Human Capabilities Credit: Sixth Tone
Text Size
- +

Toggle Dark Mode

Google’s AlphaGo, an artificially intelligent computer program, has defeated Chinese grandmaster Ke Jie at Go, a notoriously complex board game. The stunned 19-year old, who is reputedly the world’s best human Go player, remarked that AlphaGo “is like a god of a Go player” after Tuesday’s match.

The victory is yet another indication that artificial intelligence is rapidly eclipsing human capacities in a number of highly complex endeavors. Go is likely the world’s most complicated board game: a popular statistic about the ancient Chinese game is that there are more possible Go configurations than there are atoms in the universe. Even after computers surpassed humans at chess in the late 1990s, experts believed that humans would maintain dominance in Go for many more decades.

AlphaGo was developed by Google’s artificial intelligence startup DeepMind with the purpose of reaching this milestone. It originally learned how to play Go by analyzing 300 million moves made by human players. DeepMind has since developed AlphaGo to the point where it can train by playing matches against itself (a technique known as deep reinforcement learning), which means that it is continually improving.

Last year, AlphaGo trounced Korean grandmaster Lee Se-dol 4-1 in a five game series, becoming the first machine to defeat a professional Go player. AlphaGo’s triumph took humankind by surprise and was widely hailed as a breakthrough for artificial intelligence. It also changed the way humans play a game whose roots stretch back more 2,000 years, challenging long-held assumptions about Go strategy by employing unconventional tactics against Lee to devastating effect. While this demonstrates that computers can yield new insights and be used to augment human capabilities, it also underscored the fact that the gap between human and artificial intelligence will continue to grow.

Ke reportedly attempted to use some of AlphaGo’s unorthodox tactics against it during his match to no avail. “AlphaGo is improving too fast,” he said in the postgame news conference. “AlphaGo is like a different player this year compared to last year.” Ke still has two matches left against AlphaGo (which you can watch here on DeepMind’s Youtube channel), but his comments confirm that computers are now the undisputed masters of the game.

The series between Ke and AlphaGo, which is being hosted in Wuzhen, also represents a PR coup for Google given the company’s complicated relationship in China. The internet giant shuttered most of its operations in the country seven years ago due to censorship requirements.

Sponsored
Social Sharing