upper waypoint

'Like a God,' Google AI Beats Human Champ of Notoriously Complex 'Go' Game

Save ArticleSave Article
Failed to save article

Please try again

The world's top-ranked Go player, China's 19-year-old Ke Jie, reacts during the first of three matches against Google's artificial intelligence program AlphaGo on May 23. Ke Jie lost the match. (STR/AFP/Getty Images)

In the first of three matches with the world's No. 1 Go player, a Google artificial intelligence program claimed victory Tuesday. It won the round by just a fraction of a point in Wuzhen, China, but the win was enough to leave its grandmaster opponent impressed and thoroughly confounded by the result.

Last year, after Google's AlphaGo dispatched a human grandmaster in the notoriously complex board game for the first time, 19-year-old Ke Jie expressed confidence he wouldn't share the same fate, according to The New York Times. After all, Ke had defeated Lee Sedol several times before himself.

By the match's end Tuesday, Ke felt markedly different about his inhuman competitor.

"Last year, it was still quite humanlike when it played," Ke said. "But this year, it became like a god of Go."

Sponsored

His statement feels rather apt considering that, like a god, AlphaGo is its own greatest instructor. Last year, NPR's Geoff Brumfiel broke down the basics of the program developed by Google's DeepMind lab:

"It started by studying a database of about 100,000 human matches, and then continued by playing against itself millions of times.

"As it went, it reprogrammed itself and improved. This type of self-learning program is known as a neural network, and it's based on theories of how the human brain works.

"AlphaGo consists of two neural networks: The first tries to figure out the best move to play each turn, and the second evaluates who is winning the match overall."

DeepMind has only further developed the architecture of the program since it defeated Lee last year.

Also like a god, AlphaGo has been demonstrating its influence far and wide among the game's top players — including Ke, who Wired notes adopted some tactics from AlphaGo. The program played many of the world's top players under the pseudonym "Master" in online matches earlier this year, according to the magazine, and its unorthodox playing style affected Ke's.

In a blog post last year, DeepMind CEO and co-founder Demis Hassabis explained what attracted his lab to the ancient game of Go, which has been around at least since the days of Confucius:

"[A]s simple as the rules are, Go is a game of profound complexity. There are more possible positions than the number of atoms in the universe, and more than a googol times larger than chess."

It doesn't hurt that the game originated -- and remains very popular -- in China, with which Google has shared a fraught relationship. Roughly seven years ago, the California-based tech giant pulled out of the country, citing discomfort with Chinese censorship requirements.

Although Reuters reports that Eric Schmidt, chief of Google parent Alphabet, watched the match in person in China, the Times notes that "there was no obvious live video of the event" available to viewers on the Chinese mainland. Chinese officials don't allow access to Google-owned YouTube, where the match was live-streamed, or Google search.

Yet Google continues to have designs on the massive Chinese market, announcing recently that it intends to bring some of its products back to China, according to Reuters.

For now, though, skittish onlookers watching the next two matches between Ke and AlphaGo on Thursday and Saturday might do well to remember Geoff's reassurance: "This program will not lead to a dystopian future in which humanity is enslaved by killer robots.

"At least not for a few more years."

Copyright 2017 NPR. To see more, visit http://www.npr.org/.

lower waypoint
next waypoint