Advertisement

South Korean Go master retires, saying AI, machines ‘cannot be defeated’

South Korean professional Go player Lee Se-Dol says he is now retiring, after declaring that machines "cannot be defeated.".
South Korean professional Go player Lee Se-Dol says he is now retiring, after declaring that machines "cannot be defeated.". EPA/JEON HEON-KYUN/POOL

The one person who’s been able to beat Google’s algorithm at strategy board game Go has retired, saying that it was impossible to defeat machines and artificial intelligence.

According to South Korea’s Yonhap News agency, 18-time world Go champion Lee Se-dol said he decided to retire because after he realized that he was “not at the top even if I become number one.”

READ MORE: Website uses artificial intelligence to generate realistic human faces from scratch

In 2016, the South Korean national faced off against AlphaGo, an AI created by Google’s sister company Deepmind.

In a set of five matches, Lee only managed to beat the computer once.

“There is an entity that cannot be defeated,” said Lee, who is considered to be one of the greatest Go players of the modern era.

Story continues below advertisement

The 36-year-old first started playing when he was five, only turning pro seven years later.

READ MORE: Canada lacks laws to tackle problems posed by artificial intelligence: Experts

His high-profile showdown against the artificial intelligence program received both praise and concern over the capabilities of computer intelligence and machine learning.

A handout photo provided by Google shows South Korean Go grandmaster Lee Se-dol (R) making the first move in the final match of the landmark, five-game showdown between Lee and the Google-developed artificial intelligence system AlphaGo at the Four Seasons hotel in central Seoul, South Korea, 15 March 2016.
A handout photo provided by Google shows South Korean Go grandmaster Lee Se-dol (R) making the first move in the final match of the landmark, five-game showdown between Lee and the Google-developed artificial intelligence system AlphaGo at the Four Seasons hotel in central Seoul, South Korea, 15 March 2016. EPA/YNA SOUTH KOREA OUT

According to AFP, the AlphaGo has already improved to a superior, self-teaching version which has beat its predecessor in 100 games of Go to none.

Originating in China over 3,000 years ago, Go took hold as a strategy game played mostly in China, Japan and South Korea.

A.I. fuels unprecedented growth in health and science
A.I. fuels unprecedented growth in health and science

Despite the game’s simple rules, which see two players take turns placing white or black stones on a square board to try and capture the most territory, complex strategies were needed to in order attain victory.

Story continues below advertisement

According to BBC News, researchers say the game has more possible move configurations than there are atoms in the universe.

Despite Lee’s retirement, he is still scheduled to play against AI system HanDol in December.

ParityBOT uses AI to combat abusive tweets to female election candidates
ParityBOT uses AI to combat abusive tweets to female election candidates

HanDol, a program created South Korea’s NHN Entertainment Corp, has already beaten the country’s top five Go players.

Lee, who will be given a two stone advantage, expects that he will still lose.

“Even with a two-stone advantage, I feel like I will lose the first game to HanDol,” the former world champion told Yonhap News.

“These days, I don’t follow Go news. I wanted to play comfortably against HanDol as I have already retired, though I will do my best.”