The History of AI

AlphaGo beats Lee Sedol

Era 4 · Deep learning and foundation models (2012–2021) › Deep reinforcement learning

Lineage
Data and learning
Kind
Application

Who

DeepMind

What happened

AlphaGo beats Lee Sedol, one of the best Go players in the world, 4–1 in five games played at the Four Seasons Hotel in Seoul on March 9–15, 2016. Lee wins the fourth.

Why it mattered

A machine beats a top-ranked player at the board game that had held out longest against AI, before an audience of tens of millions.

The full story

Lee Sedol, a 9-dan professional, was one of the best players in the world, several ranks above Fan Hui. DeepMind challenged Lee to five games, with a $1 million prize for the winner.

AlphaGo won the first three. In the second game, its move 37, an unusual shoulder hit, took the commentators by surprise: Michael Redmond called it creative and unique. In the fourth, Lee answered with move 78, which several professionals called a "divine move," and the program lost the thread and eventually resigned. It was the only human victory. AlphaGo also won the fifth.

The match drew a huge audience in Asia: the Chinese broadcast of the first game alone reached some 60 million viewers. The Korea Baduk Association awarded AlphaGo an honorary 9 dan, and DeepMind gave the prize money to charity. In May, Google revealed that the program had run on its new TPU processors.

In May 2017 a new version of AlphaGo won all three games against Ke Jie, then the world's top-ranked player, in Wuzhen. Lee retired from professional Go in November 2019, describing AI as an entity that cannot be defeated.

Common misconceptions

  • What people often say: AlphaGo learned to play on its own, with no human games.

    In fact: That was AlphaGo Zero, in 2017. The version that beat Lee Sedol started from millions of moves from expert games and then improved by playing against itself.

Videos

Builds on

  • AlphaGo beats Fan Hui (October 5, 2015)

    Documented direct influence

    It is the same program: after beating Fan Hui, DeepMind challenged Lee Sedol with an improved version

Leads to

  • AlphaGo Zero and AlphaZero (October 19, 2017)

    Documented direct influence

    AlphaGo Zero is matched against the version that beat Lee Sedol and wins 100 games to 0 after three days of training

Threads that pass through here

Sources

How to cite

“AlphaGo beats Lee Sedol.” The History of AI, revised November 13, 2026. https://thehistoryofai.com/en/milestone/alphago-beats-lee-sedol

This entry was last reviewed on November 13, 2026.

Open in the interactive timeline