AI’s Unexpected Mastery
In March 2016, DeepMind’s AlphaGo defeated Lee Sedol, one of the world’s top Go players, 4-1 in a historic match broadcast globally. Go, an ancient Chinese board game with more possible positions than atoms in the universe, was considered decades away from AI mastery—AlphaGo’s victory shocked experts and demonstrated deep learning’s unexpected capabilities.
Move 37 & Creative AI
In Game 2, AlphaGo played Move 37—a bizarre shoulder hit that human commentators initially dismissed as an error, then recognized as genius. The move defied 2,500 years of Go wisdom, demonstrating AI discovering strategies beyond human intuition. Lee Sedol’s visible shock went viral, symbolizing humans confronting non-human intelligence.
Lee Sedol’s Response
After Game 3’s devastating loss, Lee Sedol won Game 4 with his own brilliant Move 78, proving human creativity still mattered. His retirement from professional Go in 2019 cited AlphaGo specifically: “Even if I become number one, there is an entity that cannot be defeated.” The existential weight of competing against unchallengeable AI haunted his decision.
AI’s Leap Forward
AlphaGo used neural networks trained on 30 million human games, then self-play reinforcement learning. AlphaGo Zero (2017) achieved superhuman performance learning only from self-play—no human knowledge required. The victory accelerated AI investment globally and raised questions about human exceptionalism, the nature of creativity, and employment futures in fields requiring pattern recognition.
Sources:
- DeepMind Paper (Nature): https://www.nature.com/articles/nature16961
- Documentary “AlphaGo”: https://www.alphagomovie.com/
- IEEE Spectrum coverage: https://spectrum.ieee.org/alphago-wins-match-against-top-go-player