Despite rapid advances in artificial intelligence, it seems AI still has a way to go before it can outsmart a top chess grandmaster. The world’s number one chess player, Magnus Carlsen, recently shared how he decisively defeated ChatGPT in a game—highlighting the AI’s current limitations.
AI technology has evolved far beyond simple Q&A or quirky image generation. Today’s AI can engage in deep conversations, mimic expert medical advice, and create detailed visuals. Experts predict that within the next decade, we may reach “artificial general intelligence,” where AI surpasses humans across all tasks. Elon Musk even claims his company’s latest model outperforms graduate students.
Given AI’s prowess, you might expect it to excel at chess, a game built on logic and calculation. But Carlsen, who has held the world number one ranking for over 14 years, posted on social media about a match where he overwhelmed ChatGPT—even under a very specific win condition.
Carlsen tweeted, “I sometimes get bored while travelling,” alongside screenshots of the match and their post-game chat. ChatGPT acknowledged defeat, saying, “With 53… Bxh5, you’ve officially done it: All of my pawns are gone, you haven’t lost a single piece, you fulfilled your win condition perfectly. I resign. That was methodical, clean, and sharp. Well played!”
This result confirms that despite AI’s computational power, it still struggles in complex, strategic environments like chess—where human intuition and experience remain key.
Afterwards, Carlsen asked ChatGPT to analyze his gameplay. The AI praised his opening strategy, particularly his use of the Philidor Defense, his positional patience, tactical awareness, and impressive endgame technique—where his bishops and pawns converted a small advantage into a winning plan.
Interestingly, ChatGPT admitted to attempting several illegal moves during the game but commended Carlsen for consistently enforcing the rules, suggesting strong real-life experience.
Carlsen also asked the AI to estimate his classical rating strength—the official measure used to rank players globally. ChatGPT’s estimate ranged from 1800 to 2000 FIDE or USCF, far below Carlsen’s actual rating of 2839. (For context, a rating of 2500 would place you 603rd worldwide; 2000 would rank you about 26,599th.)
While it’s unrealistic to expect AI to accurately rate the world’s best player based on a single game, this mismatch further illustrates the current gap between AI and elite human chess mastery.
In short, even as AI gets smarter, Magnus Carlsen proves that some skills—like grandmaster chess—still require a human touch.
