Shin Jin-seo competes against Byun Sang-il in the sixth Sopalkosanol Top Player Championship at K-Baduk Studio in Seongnam, Gyeonggi, on June 25. Shin won by resignation as White in 206 moves and swept the championship 3-0.KOREA BADUK ASSOCIATION
Shin Jin-seo, the world's top-ranked Go player, says he is aiming for at least two wins against the AI program KataGo in the first showdown between a human and a Go AI in a decade.
The series is the first major human-versus-AI meeting in Go, the board game known as Baduk in Korean, since Lee Se-dol lost to Google's AlphaGo in 2016. Shin will play with a two-stone handicap across three games, the first on Friday and the next two on Sunday and next Tuesday.
Shin sounded confident about his chances of winning.
"In practice, I want to aim for two wins, or even three," he said during a press conference held Tuesday at the Korea Baduk Association in Seongdong District, eastern Seoul. "The key will be how well I respond to KataGo's moves in the middle game."
KataGo is the most widely used open-source Go AI since AlphaGo, used by professional players and national team members for study and training, as well as by television commentators.
Shin sees the game turning on how long he can keep it out of a direct fight.
"If it turns into a fighting game, my chances are below 10 percent, but if I can take it into the endgame, I'd put them at around 60 to 70 percent," he said. "I'll focus on shaping the board so the game is settled in the endgame, in my own style."
Shin will receive a match fee of 50 million won ($33,400) per game, or 150 million won in all, and a 50 million won bonus for each win. Two or more wins would also earn him a Genesis G90 sedan.
Shin Jin-seo is interviewed with Demis Hassabis, CEO of Google DeepMind at the AlphaGo 10 Years: Great Partnership event held at Westin Chosun Hotel in central Seoul on April 29.WOO SANG-JO
Ahead of the series, Shin has trained intensively against an AI of similar strength to KataGo.
"At first, I could not win a single game at a two-stone handicap, but as I prepared to win after the offer came, I began winning more," he said. "A two-stone handicap is well worth challenging."
The contest, he said, is not the one it was a decade ago.
"In the AlphaGo era, you could lure the AI into mistakes with clever moves, but now you have to meet it with skill," Shin said. "Thanks to AI, opening and endgame play have improved enormously, but the game is still decided mid-game."
For Shin, the result is not the only measure.
"Win or lose, this match will be a great help in my future games against human players," he said. "I'll give my all. I am determined to show everything about my Go."
This article was originally written in Korean and translated by a bilingual reporter with the help of generative AI tools. It was then edited by a native English-speaking editor. All AI-assisted translations are reviewed and refined by our newsroom.
