Impact of Go AI on the professional Go world
photo credit: Dan Maas
I originally prepared this note for the AI Roundtable, a program produced for the opening ceremony of the 2020 US e-Go Congress. (AI Roundtable video link: https://youtu.be/-cEL7I6BWTc?t=3307)
Professional Go players are often referred to as one of two types, tournament players or teaching players. Tournament players are the ones who spend most of their time and energy training and competing in tournaments, and most of their income is tournament prize money. Teaching players may still compete in tournaments, but they have other economic activities such as commentating on TV, writing Go books, running Go schools, offering Go lessons, and so on.
In the case of Korea, there are about 380 professional players certified by the Korean Baduk Association, and about 50 top players can be considered tournament players. Last year, the top player in Korea earned about $1 million US dollars from tournaments alone, while the 10th player earned about $120,000 US dollars.
The development of superhuman Go AI has impacted the professional Go world in many different ways. Here, I want to highlight three areas, one that affects both types of players, one area for tournament players, and one for teaching players.
Want to read this story later? Save it in Journal.
The first area is somewhat abstract, but it affects all current and future professional players — Go as a career or life path. When I was an aspiring pro, I lived at my Go master’s place with the Go master’s family and several other fellow students. Once in a while after dinner, my Go master would call all his students to sit around the tea table in the living room and calmly make a pot of green tea following all the proper rituals. Often he would begin with some tips about tea leaves or tea table manners. Then he would ask us how we are doing or whether we had any new thoughts about Go.
At the time, which is not a very long time ago, no one questioned that Go was a path you walk for a lifetime. The life of a Go player was considered similar to a philosopher, a scholar, an artist, or a monk. Although there was a new movement trying to convince people that Go is actually a mind sport and Go players are elite athletes, many Go players harbored more traditional values inside them. The belief was, as a professional player, you explore and endeavor to reach an ever higher level of understanding.
The term “divine move” is used as a metaphor for an ultimate level of play. With AI, however, we all realized that the best way to reach the highest possible level of Go is not through thinking about it for a lifetime. It’s actually to buy more powerful GPUs and a well-trained deep neural network and have it play Go. So, suddenly, we players felt an enormous sense of loss. When AlphaGo defeated Lee Sedol in the first match, even though I had already changed my career path, I still felt that loss inside me. Different pros are responding to it differently, but many of them are choosing to leave Go. The case in point is Lee Sedol, who retired at the end of last year. In a recent interview on TV, he described this loss of mission as a major reason for his retirement.
The second important change I see is the professional players’ race to learn from AI. When I visited Korea last December, I had a chance to catch up with a few of my pro friends. They told me that in the pro circuit, it’s a unanimous belief that you need to play like an AI to win, and every serious player has an AI set up at home. Now, there are both up and down sides to this. The upside is that we sometimes see a player who was somewhat past his prime suddenly climb back to the top, having trained with AI more intensely. There are a growing number of young and new pros who demonstrate surprising strength. This change gives hope to all pros who dream to become number one, and also makes competitions more interesting to fans as well. On the downside, however, pro players have lost the passion or motivation to develop their own styles. Before AI, most strong players had distinctive flavors of play and oftentimes this style was the reason why some Go fans rooted for one player over the other. Today, everyone is trying to imitate the AI style, and the pros judge each other only by who is better at playing like the AI.
The third area I want to discuss affects teaching pros, and it saddens me a bit. The demand for pro-level teaching games and private lessons has plummeted. Professional players used to command a high price for teaching games and lessons, and this has been a critical source of income for many pros. Typical clients are parents of young students who are studying Go competitively. Usually how it works is that a headmaster of Go school asks the parents of students if they want additional training for their child. If the parents can afford this additional training cost, the Go school recruits a professional player, who would agree to play a fixed number of games, like five games, one per week, at a certain price. Then the pro comes to the Go school, plays a teaching game and reviews it with the student afterward. However, high level games and reviews can be replaced by a strong AI now. Of course there is still room for lower level classes and teaching, but pros are often better at playing teaching games than explaining easy concepts. This resulted in many pros struggling to find a new livelihood.
These are the three major impacts of Go AI in the professional Go world that I have observed. Now Go can be seen more as a mind sport than a life pursuit. Students and Go fans have gained powerful AI tools to help them understand pro games and learn to play better. But pros who have supported themselves with a teaching career are struggling to find a new direction.