Future Tense

China’s Best Go Player Lost a Game to an A.I. The Chinese Government Censored It.

People in China did not get to watch an A.I. take down a champion Go player.


There are trends, and then there are symbols: narrative inflection points that speak volumes. Tuesday became one of the latter when an A.I. called AlphaGo defeated Chinese national Ke Jie, the world’s top-ranked player in the ancient Chinese strategy game Go. The program narrowly bested Ke in the first of a series of engagements that lasts until Saturday. Held in Wuzhen, China, the matches are taking place under Google’s auspices as part of a conference called the Future of Go Summit.

But the man-vs.-machine battle was barred from view in China. As China Digital Times reported Wednesday, the Chinese government issued censorship reports decreeing that “this match may not be broadcast live in any form and without exception, including text commentary, photography, video streams, self-media accounts and so on. No website (including sports and technology channels) or desktop or mobile apps may issue news alerts or push notifications about the course or result of the match.”

Why was China so bearish on Ke’s chances? One answer is self-esteem. As the Guardian’s Alex Hern reported Wednesday, all signs point to the government’s fears that its 19-year-old world champion losing on his home turf “would hurt the national pride of a state which holds Go close to its heart.” Hong Kong–based New York Times technology reporter Paul Mozur articulated another possibility: envy. AlphaGo is a product of DeepMind Technologies, a London-based startup founded in 2010. Four years later, Google—based, of course, in the United States’ Silicon Valley—bought DeepMind for a reported $625 million. AlphaGo’s “victory took place in China, a rising power in the field of artificial intelligence that is increasingly seen as a rival to the United States,” Mozur wrote.

But defeating Ke was hardly AlphaGo’s first rodeo. In February 2015, a cover article in the science journal Nature co-written by DeepMind’s co-founder, Demis Hassabis, announced that the company’s A.I. algorithm, christened the Deep-Q Network, had managed to “achieve a level comparable to that of a professional human games tester across a set of 49 games” including Atari’s Boxing, Pong, and Space Invaders. Less than a year later, a second Nature article touted an even greater milestone: DeepMind’s technology, now branded AlphaGo, had both mastered Go and defeated its thrice-over European champion, France’s Fan Hui, five games to zero in a secret matchup months earlier—the first time an A.I. machine had ever done so without assistance.

Go, today played across China, Japan, and Korea, dates from at least the 6th century BCE. The game is dizzyingly complex with more possible positions for the black-and-white game pieces—called stones—than there are atoms in the known universe. (Seriously.) “AlphaGo plays in a very human style, because it’s learned in a human way and then got stronger and stronger by playing, just as you or I would do,” Hassabis, himself a former chess prodigy, told the Guardian last year.

And AlphaGo didn’t stop there. In March 2016, it faced off against Lee Sedol, an internationally ranked Go player from South Korea. Although experts (including Lee himself) were bullish on a human win, AlphaGo dispatched the grandmaster with relative ease. Then, over a period of several days culminating in early January, AlphaGo defeated scores of international Go experts in 60 anonymous online games.

Expectations ran high ahead of Tuesday’s match against Ke. As Quartz’s Zheping Huang phrased the stakes on Tuesday, “the 19-year-old reigning top-ranked player is fighting against complete machine dominance over what is possibly the world’s most complex board game.” After AlphaGo’s January online spree, Ke boasted on Weibo, the Chinese version of Twitter, that he still had “one last move” in mind to compete with the A.I.’s exploits. He lost his first showdown against AlphaGo by a slim half-point.

The program’s victorious tear may have also cost it that human style of play Hassabis raved about. After his defeat, Ke noted, “Last year, it was still quite human-like when it played. But this year, it became like a god of Go.” The Future of Go exhibition matches will also see AlphaGo partner up with a flesh-and-blood sidekick to take on two human players simultaneously and attempt to hold its own against five Go masters working together to unseat the A.I. kingpin, the Verge reported Monday. Ke still has two more chances to defend humanity’s honor from the encroaching A.I. juggernaut.

After Lee’s loss, former Jeopardy! champion Ken Jennings—who famously lost to IBM’s Watson A.I. in 2011—wrote in Slate, “in a very real way, your opponent isn’t just a room full of servers or a few thousand lines of code. It’s the Future, the possibility that your own individual talent, the thing that’s made you special your whole life, can now be replaced by a sufficiently clever algorithm.” Whether or not Ke’s able to turn the tide against AlphaGo later this week, the Chinese people likely won’t be watching as that future continues to bear down on us.