AI reshapes Go's top players' thinking

AI now trains Go's best—humans follow the machine.

Inside a quiet building in Seoul’s Hongik-dong, the hush of a traditional Go academy has given way to a different rhythm: screens glow, clicks echo, and players replay games with machine-perfect precision. The Korea Baduk Association, long the temple of elite strategy, has become a hub where centuries-old tacticians and cutting-edge AI sit side by side. Ten years after AlphaGo stunned the world by beating Lee Sedol, AI is no longer a novelty; it’s the core of how the game is learned, practiced, and even judged.

The shift is stark: competitors no longer rely on intuition alone. They study thousands of AI-replayed lines, compare positions with the machine’s verdicts, and debate the AI’s preferred routes long after a game ends. The point is not merely to imitate a best move but to align with the machine’s overall way of thinking—a form of training that reshapes what professionals consider “the right move” in a given position. “Essentially impossible to compete professionally without using AI,” players and coaches are telling one another, not about a trend but a baseline.

The transformation is not just about speed or accuracy. It’s about a redefinition of creativity itself. Pro players once prided themselves on inventing lines that surprised opponents. Now, many aim to reproduce or approximate AI’s ideas, even when the machine’s reasoning remains opaque to human observers. Some critics argue that this erodes the game’s human spark; others see it as a new kind of ingenuity—the art of translating machine insight into human play.

One enduring consequence: broader access. If AI can dissect positions with the patience of a digital mentor, training becomes less dependent on one-on-one mentorship from a living grandmaster or a single, expensive coach. The technology is democratizing practice—an effect some analysts say is already visible in the sport’s evolving hierarchy, where more female players are climbing toward the upper echelons. The game’s old gatekeepers—geography, funding, and networking—now share the stage with AI-powered coaching tools that can be deployed in smaller academies and even online.

For Shin Jin-seo, currently ranked at the top of the Go world, AI is a constant companion in practice—an “invaluable training partner,” as insiders put it. The extent of AI’s influence is not just about moves; it’s about a shared language of position evaluation, risk assessment, and strategic tempo that players must adopt to stay competitive in a field where the machine’s preferences shape the expected outcomes of almost every major match.

From a practitioner’s lens, the Go AI revolution is instructive for teams in other high-skill domains. First, access matters more than raw capability: once AI analysis becomes a baseline service, the winners are those who make it affordable and easy to integrate into daily coaching. Second, there’s a risk of homogenization. When thousands of players converge on AI-recommended lines, the space for unconventional, human-driven invention narrows unless coaches insist on preserving it as a deliberate practice. Third, evaluation and provenance matter. If AI becomes the primary benchmark, teams must ensure that the tools used capture genuine improvement and don’t merely teach players how to imitate a machine’s phrasing.

What this means for products shipping this quarter is tangible. Expect AI-assisted coaching tools to migrate from niche research labs into mainstream Go academies and consumer apps. Training platforms will offer real-time AI analysis of positions, automated post-game reviews, and dashboards that track a player’s evolution relative to AI-curated “best” lines. Sponsors and associations will increasingly fund compute-supported practice programs, expanding access for rising talents who previously faced barriers to elite-level study. In short: smaller, cheaper, better training is here, powered by the same force that changed AlphaGo’s century-old rival—AI.

The core idea is simple, even jolting: AI is not just augmenting human play; it’s rewriting how the world’s best Go players think, one move at a time. The result is a game that moves faster, with new ideas that can feel familiar to the machine and, increasingly, familiar to the human watching it.

Sources

AI is rewiring how the world’s best Go players think