Lee Seidol was the best Go player of his generation when he suffered a decisive loss, beaten not by a human opponent but by artificial intelligence.
Mr Lee was defeated by AlphaGo, an AI computer program developed by Google's DeepMind unit. The astonishing upset, in 2016, made headlines around the world and seemed like a clear sign that artificial intelligence was entering a new, deeply troubling era.
Credited to 18-time world champion Mr. Lee for his intuitive and creative style of play, AlphaGo solves one of computer science's greatest challenges: teaching itself the abstract strategies needed to win at Go. , which is widely considered to be the most complex board in the world. the game
“I am very surprised because I never thought I would lose,” Mr Lee said at a post-match news conference. “I didn't know AlphaGo would play such perfect Go.”
But the implications of his loss went far beyond the game itself, in which two players compete for territory by placing black and white stones on a 19-by-19 grid board. AlphaGo's victory demonstrated AI's unbridled ability to achieve superhuman mastery of skills once considered too complex for machines.
Mr. Lee, now 41, retired three years later, convinced that humans could no longer compete with computers at Go. He said that artificial intelligence has changed the nature of the game that started in China 2500 years ago.
“Losing to AI meant my whole world was falling apart,” he said in a recent interview with The New York Times.
As society grapples with AI for the future of humanity, Mr. Lee is now urging others to avoid being caught unprepared, as he was, and to become familiar with the technology now. He lectures about AI, trying to give the others the advance notice he wishes he had received before his match.
“I encountered problems with AI early on, but it will happen to others,” Mr. Lee told a crowd of students and parents at a community education fair in Seoul recently. “It can't be a happy ending.”
Since his loss, Mr. Lee has become something of an AI fanatic, with a frenzied focus as artificial intelligence delivers one breakthrough after another.
AI has helped chatbots carry on conversations that are almost indistinguishable from human interaction. It has solved problems that have puzzled scientists for decades, such as predicting protein conformations. And it has blurred the lines of creativity: writing music, producing art and making videos.
Mr. Lee is not a Dom Seer. In his view, AI may replace some jobs, but it can also create anything. Reflecting on AI's grip on Go, he said it's important to remember that humans created the game and designed the AI systems that mastered it.
What he worries about is that AI could replace the value of humans.
“People used to be in awe of creativity, originality and innovation,” he said. “But since AI came along, a lot of that has disappeared.”
Mr Lee started playing Go at the age of 5 under the guidance of his father, a school teacher and sports enthusiast. His family lived on Bigeumdo, an island off the southwest coast of the Korean Peninsula with a population of about 3,600.
His immense abilities were evident from the beginning. He soon became the best player of his age not only locally but throughout South Korea, Japan and China. He turned pro at the age of 12.
By the time he was 20, Mr. Li had reached 9-dan, the highest level of proficiency in Gu. Soon, he was among the best players in the world, dubbed by some as the Roger Federer of the game.
“He was an idol, he was a star,” said Lee Hajin, a former professional Go player. “Everybody looked up to him,” Ms Lee added
As Mr. Li's stature grew, Gu began to draw interest from a new audience: computer scientists.
Go presents a daunting challenge for AI researchers. The game is exponentially more complex than chess, with it often said that there are more possible positions on the go board than there are atoms in the universe (followed by more than 100 zeros, according to many mathematical calculations). ).
The breakthrough came from DeepMind, which built AlphaGo using so-called neural networks: mathematical systems that can learn skills by analyzing huge amounts of data. It started by feeding the network 30 million moves from top-level players. The program then played game after game against itself until it learned which moves were successful and developed new strategies.
By the end of 2015, AlphaGo had defeated the three-time European Go champion five times in a row in a closed-room match.
Subsequently, Mr. Lee was approached by Ms. Lee, a former professional working at the International Go Federation, with a proposal for a public match, with a $1 million prize for beating AlphaGo.
Mr Lee said he accepted the offer without thinking, thinking it would be a “joke”.
“But it's fun with the assumption that I'm going to win,” he said. “The possibility of losing did not occur to me.”
The best-of-five match played in Seoul was a spectacle. In South Korea, where millions of people play Go and Mr. Lee is a celebrity, the showdown led the night's television broadcasts. More than 200 million people watched, with large audiences in China and Japan.
During the matches, a DeepMind engineer sat next to Mr. Lee and placed the stones as they were assigned to him by AlphaGo. Not having a real human rival is troubling, Mr Lee said. AlphaGo played a style he had never seen before, and it felt strange not trying to understand what his opponent was thinking and feeling. The world watched in amazement as AlphaGo cornered Mr. Lee and made moves unimaginable to a human player.
“I couldn't get used to it,” he said. “I thought AI would beat humans someday. I didn't think it was here yet.
AlphaGo won 4 out of 5 matches. Lee Sang-hoon, his older brother and a professional Go player, remembers thinking: “This can't be.”
“It was shocking,” said his brother, who continues to play as a professional. Like other professionals, he now trains with AI systems that continue to learn and improve.
“Pro players are studying how these algorithms work and trying to close the gap,” his brother said. “But we're a long way off.”
DeepMind Chief Executive Damis Hassabis said in a written statement that AlphaGo's victory was a defining moment in the history of AI. It showed that computers that learn on their own from data “were really worth it,” he said.
Mr Lee found it difficult to accept defeat. What he considered an art form, an extension of a player's own personality and style, was now set aside for the ruthless efficiency of algorithms.
“I can't enjoy the game anymore,” he said. “So I retired.”
Mr. Lee has taken a step into the Go world. He has written several books, including an autobiography and a series about his famous matches. He has created board games inspired by Go. He founded the Go Academy for Children, which has about a dozen branches across the country.
But AI dominates his thoughts, partly because of the ambivalence he feels about the pros and cons, but also because it's a topic that hits close to home.
Her 17-year-old daughter is in her final year of high school. When they discuss what he should study at university, they often consider a future shaped by AI.
“We often talk about choosing a job that cannot be easily replaced by AI or is less affected by AI,” he said. “It's only a matter of time before AI is ubiquitous.”
Cadet Metz Contributed reporting from San Francisco.