AIster · Prologue · Events

AlphaGo beats Lee Sedol at Go

An AI beat one of the best Go players in the world, with a move few humans would play

Key facts

Date

9–15 Mar 2016

Era

Prologue

Interest

Events

Threads

Model wars, Labs & people

Who

DeepMind's AlphaGo v Lee Sedol

Where

Four Seasons Hotel, Seoul

Result

AlphaGo 4, Lee Sedol 1

Prize

$1 million, donated to UNICEF, STEM and Go charities

The match

Between 9 and 15 March 2016, DeepMind's program AlphaGo played Lee Sedol in a five-game match of Go at the Four Seasons Hotel in Seoul, South Korea. Lee, winner of 18 world titles, was widely seen as the best player of the decade. AlphaGo won 4–1.

Go is an ancient board game in which two players place black and white stones to surround territory. It had long been treated as a much harder target for computers than chess, which IBM's Deep Blue had conquered in 1997.

Move 37

The moment people remember came in game two. AlphaGo's 37th move was one that, by DeepMind's account, had only a 1 in 10,000 chance of being played. Professional commentators found it strange, and Lee took an unusually long time to respond. Lee later said: “I thought AlphaGo was based on probability calculation and that it merely was a machine. But when I saw this move, I changed my mind. Surely, AlphaGo is creative.”

Lee's win

Lee took game four on 13 March. His 78th move, which DeepMind says also had a 1 in 10,000 chance of being played, put AlphaGo under pressure it could not recover from. It was the only game he won in the match.

Prize, rank and audience

The winner's prize was $1 million. Because AlphaGo won, DeepMind gave the money to UNICEF and to charities supporting STEM education and Go. AlphaGo was also awarded a 9 dan professional rank, the highest in the game.

DeepMind says over 200 million people watched the match worldwide.

Why it matters

AlphaGo showed that learning systems could master a task that many experts had expected to resist computers for years longer. It also changed the public conversation. A machine had produced an idea that surprised the best human players.

Questions people ask

When did AlphaGo beat Lee Sedol?

AlphaGo beat Lee Sedol 4–1 in a five-game match played in Seoul between 9 and 15 March 2016. Lee won game four on 13 March.

What was AlphaGo's move 37?

Move 37 was AlphaGo's unusual play in game two against Lee Sedol. DeepMind says it had a 1 in 10,000 chance of being played, and Lee said afterwards that it changed his mind about whether AlphaGo was creative.

Sources

1. Google DeepMind, AlphaGo, 2026-10-01
2. Google, AlphaGo’s ultimate challenge: a five-game match against the legendary Lee Sedol, 2016-03-08
3. TechCrunch, Defeated Go world champion beats DeepMind AI in penultimate match, 2016-03-13