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

2026-02-27 · Source: MIT Technology Review · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Gaming & Interactive Media · Depth: Novice, medium

Summary

Ten years after Google DeepMind's AlphaGo defeated Lee Sedol, AI has fundamentally reshaped professional Go, overturning centuries-old principles and introducing new strategies. Players now extensively use AI programs like KataGo, nicknamed "Shintelligence" by top-ranked Shin Jin-seo, to analyze moves and train, with Shin's moves matching AI's 37.5% of the time in 2022. This shift has led to a homogenization of opening moves, with many players memorizing AI-suggested sequences, and the game's crux moving to complex middle-game calculations. While some lament a loss of creativity, AI has also democratized training access, particularly benefiting female players like Kim Chae-young, who have achieved new milestones. Researchers are actively trying to extract and codify the "superhuman knowledge" from AI systems like AlphaZero to make its reasoning more transparent to human players.

Key takeaway

For professional Go players or coaches seeking to maintain competitiveness, integrating AI tools like KataGo into your daily training is no longer optional. Focus on understanding the "why" behind AI's suggested moves, even if its reasoning remains opaque, to develop a new intuition. This approach is crucial for adapting to the evolving meta-game and potentially uncovering new strategic paradigms, despite concerns about homogenization.

Key insights

AI has transformed professional Go, shifting training paradigms and democratizing access while raising questions about creativity and human understanding.

Principles

• AI can overturn established domain heuristics.
• AI training can homogenize human performance.
• AI can democratize access to high-level training.

Method

AI Go programs like AlphaGo and KataGo are trained on vast human move data or through self-play, learning to predict winning probabilities and optimize scores across the entire board.

In practice

• Use AI to identify optimal opening moves.
• Analyze AI's "blue spot" suggestions for strategy.
• Re-evaluate traditional domain intuitions against AI.

Topics

• AlphaGo
• KataGo
• Game AI
• AI Training
• AI Explainability

Best for: AI Scientist, Research Scientist, AI Researcher, Domain Expert, General Interest

Related on AIssential

• From games to biology and beyond: 10 years of AlphaGo’s impact — AlphaGo's victory in Go catalyzed modern AI, demonstrating systems can surpass human intuition and accelerate scientific discovery.
• How Tech Changed Chess — AI, despite surpassing human chess ability, has paradoxically expanded human strategic imagination and sustained the game's cultural relevance.
• Building a Chess Coach — Anant Dole and Asbjorn Steinskog, Take Take Take — Combining classical chess engines with LLMs and human-like move prediction enhances AI chess coaching.
• WallZero: Mastering the Game of WallGo with Strategic Analysis — WallZero, an AlphaZero variant, demonstrates superior WallGo play and can analyze game fairness and optimal strategies.
• MMG2Skill: Can Agents Distill In-the-Wild Guides into Self-Evolving Skills? — MMG2Skill enables agents to distill and self-evolve executable skills from diverse, real-world human guides, improving task performance.
• [P] Vibecoded on a home PC: building a ~2700 Elo browser-playable neural chess engine with a Karpathy-inspired AI-assisted research loop — An AI-assisted research loop can rapidly develop high-performance neural chess engines on consumer hardware.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by MIT Technology Review.