The 2016 Go match between AlphaGo and Lee Sedol stands as a watershed moment in artificial intelligence history. The defeat of Lee, an 14-time world champion considered one of the greatest Go players ever, marked the first time an AI had mastered a game once thought to require unique human intuition and creativity. Unlike chess, which Deep Blue conquered through “brute force” calculation in 1997, Go’s vast possibility space and reliance on pattern recognition and positional judgment had led many experts to predict it would take decades for computers to outperform humans.
AlphaGo’s victory, however, achieved through deep learning techniques and reinforcement learning, demonstrated that AI could now operate in domains previously thought to be exclusively human. The match’s dramatic five games, watched by over 200 million viewers worldwide, crystallized for many the realization that artificial intelligence had crossed a fundamental threshold.
Eight years later, at a special lecture on “Artificial Intelligence and the Future of Creativity” hosted by Seoul National University’s ELSI Research Center for Artificial Intelligence, Lee Sedol looked back on this transformative period with unique insight. As we now grapple with AI’s impact on creative domains like music, writing, and art, the game of Go offers a prescient case study — it represents a field that has already undergone a complete AI-driven transformation.
The Old Vision of Go
Lee Sedol announced an early retirement in 2019 at the age of 36, three years after his famous AlphaGo match. After retirement, Lee debuted as a board game designer but has maintained a thoughtful and often critical voice in Go discourse.
Among the most poignant moments of the ELSI lecture was Lee’s reflection on Go as an art form that transcends mere competition. For him, the game’s deepest beauty emerges from the collaborative dialogue between two players. He went on to speak movingly of his unfulfilled dream of creating a true “masterpiece” in Go history — not only through individual brilliance, but through the rare alchemy that occurs when two players engage with total concentration and, most importantly, mutual respect. According to his philosophy, Go is a unique form of two-person artistic creation, where meaning arises from the exchange of intentions and the subtle rapport that develops over a game’s duration. Winning and losing, he maintained, were always secondary concerns as Go contains no absolute answers, like any profound art form — instead its true value lies in how players navigate uncertainty and forge their own paths, potentially inspiring other Go learners along the way.
Perhaps that’s why Lee’s relationship with AI seemed complex and tinged with melancholy. This was most evident when he was recounting the famous fourth match—his sole victory against AlphaGo and humanity’s last win against AI in professional Go. (Shortly after Lee’s match, then world #1 Ke Jie faced AlphaGo and was defeated comprehensively. AlphaGo retired after this match, though in subsequent years, some amateurs have managed occasional wins against other AI system.) Paradoxically, this achievement became a source of inner conflict because the victory, achieved by deliberately exploiting a potential bug in AlphaGo’s programming, contradicted his artistic philosophy of the game. The match’s singular focus on winning, divorced from the beautiful interplay he valued between two minds, eventually led to his growing disillusionment with the direction of professional Go in an AI-dominated era. Reflecting back on his early retirement and Go’s trajectory after the AlphaGo match, he even suggested that “I could not recognize it — it was not the same Go that I used to know.”
So, how has this ancient game, one that has stood the test of time for over 4,000 years, been so fundamentally transformed in less than a decade?
The Transformation
In short, Go has undergone a radical transformation that runs counter to Lee’s artistic vision. AI programs have become ubiquitous learning tools and a singular source of inspiration for Go learners, dramatically changing how both professionals and amateurs understand and approach the game. The traditional barriers to entry — which once required years of dedicated study and practice — have been dramatically lowered. Players can now achieve considerable skill levels in relatively short periods by mechanically memorizing AI-recommended moves.
The current world champion, Shin Jin-Seo, epitomizes a new generation of players who learned Go primarily through AI rather than traditional training methods. While previous generations honed their skills in private Go academies, studying the classic games of past masters along with their peers, Shin’s generation has been shaped by AI from the start. Interestingly, professional rankings now correlate strongly with how closely players can match AI move predictions. This transformation is perfectly captured in Shin’s nickname, ‘신공지능’ (Shin Gong Jineung)—a clever wordplay combining his surname (Shin) with”’artificial intelligence” in Korean. His remarkable achievement of scoring 92% in AI move prediction accuracy underscores a fundamental shift in how excellence in Go is measured. The more closely you resemble AI, the more likely you are to win.
This transformation sparks debate among the Go communities about whether Go is now what Lee Sedol terms a “mind sport” — more focused on finding the right answer than cultivating creative expression. While AI can suggest statistically optimal moves, it cannot explain its reasoning, leading to a form of rote learning that lacks the deep reflection on intention that characterized traditional Go. Perhaps more importantly, the ability to achieve high-level play through AI assistance, without understanding the underlying principles or developing a personal style, raises fundamental questions about the nature of mastery itself.
Broader Implications
The pattern is becoming familiar: AI dramatically lowers barriers to entry, making previously elite skills widely accessible. We now have the ability to achieve results without going through the process or investing time—whether in Go, music, writing, or design. It’s more convenient, faster, more productive. Some argue this is simply human evolution, that these machines are natural extensions of our capabilities, just another chapter in the story of progress.
But we might ask ourselves: suppose you can now casually defeat the world’s best Go player using AI, just as you are now able to generate a thousand novels in a day using a genAI tool — and one of them might even win a literary award. What does that achievement really mean? More importantly, what does it tell us about ourselves?
Philosopher Shai Tubali recently highlighted a profound observation from Indian philosopher Jiddu Krishnamurti regarding “the mechanized mind”:“While a computational functionalist might argue that building a mind is as simple as building a machine, Krishnamurti believed our minds are so much more—he worried we were selling ourselves short, letting our minds get stuck in mechanical routines like memory and knowledge-crunching (…) So, the real question isn’t whether AI will develop human-like minds — it’s whether we’re slipping into machine-like minds ourselves (…) Consciousness itself doesn’t set us apart if we reduce the mind to computation.”
Technologies like AI will reshape parts of our human experience and redefine who we are, as evidenced by Go’s transformation where values of efficiency and optimization now prevail. This shift indeed echoes Krishnamurti’s observation about the “mechanized mind.” But perhaps in accepting this transformation so readily, we’re gazing at reflections of our mechanical selves, voluntarily reducing the rich tapestry of human experience to machine-like patterns. As Krishnamurti suggests, we may be “selling ourselves short.”
As we find ourselves in this pivotal moment of technological change, we face a delicate balance. While some traditional human experiences will naturally give way to automation, we retain the power to protect those elements we deem essential to our flourishing. The challenge before us is to proactively choose which aspects of human experience—whether the patient joy of mastery or the rich texture of imperfect creation — we refuse to sacrifice at the altar of relentless optimization. In this light, Go’s transformation isn’t just a cautionary tale; it’s an invitation to reflect seriously on what’s at stake: that is, what makes us human, and what we must fight to preserve.
All of which brings me back to my childhood memories—I too learned Go between the ages of 8-12 in the traditional way, and although I don’t remember much of how to play the game anymore or all its intricacies, what stays in me are those quiet hours spent alone in my room, struggling with the old masters’ plays. Looking back, I see these as the kind of non-mechanical, deeply human experience Krishnamurti (and Lee Sedol to an extent) advocated—where meaning emerges from the patient struggle to understand something beyond ourselves. Such moments feel increasingly rare in my life now, where experience often defaults to the very mechanical routines Krishnamurti feared—my mind willingly reducing itself to nothing more than an input-output machine, seeking the path of least resistance.
This is what hits home so strongly when I think about what might be lost in our rush toward a mechanized mind. And it is memories like this that compel me to respond to the challenge of AI—to protect those spaces where my mind can unfold at its own pace, even as we’re told that the boundaries between our minds and machines grow increasingly fluid. What experiences drive you to do the same?
Joseph Park is the Content Lead at DAL (joseph@dalab.xyz)
Illustration: Soryung Seo
Edits: Janine Liberty
