The Magic of Move 37: How AlphaGo Redefined Creativity and Intelligence

Watch the documentary here: https://www.youtube.com/watch?v=WXuK6gekU1Y

Introduction

In March 2016, the world witnessed a pivotal moment in artificial intelligence history: AlphaGo, an AI developed by DeepMind, defeated Lee Sedol, one of the greatest Go players of all time, in a five-game match. While the entire series was groundbreaking, one move in the second game—Move 37—captured global attention. This move, initially dismissed as a mistake by human experts, revealed a level of creativity and strategic depth that forever changed our understanding of AI and human intuition. In this blog post, we explore the story behind Move 37, its implications for AI, and the lessons it holds for human creativity and innovation.

Understanding the Challenge of Go

Go is an ancient board game that originated in China over 2,500 years ago. Despite its simple rules, it is immensely complex due to the vast number of possible moves and the intuitive, strategic thinking required. Unlike chess, where computers like Deep Blue succeeded by brute-force calculation, Go was long considered a “holy grail” for AI researchers. As Demis Hassabis, CEO of DeepMind, noted:

“Go is incredibly challenging for computers to tackle because compared to, say, chess, the number of possible moves in a position is much larger. In chess, it’s about 20. In Go, it’s about 200. And the number of possible configurations of the board is more than the number of atoms in the universe.”

This complexity meant that traditional AI methods, which relied on pre-programmed rules and heuristic searches, were insufficient. Mastering Go required something more: intuition and creativity.

The Rise of AlphaGo

AlphaGo, developed by DeepMind, took a revolutionary approach. Instead of relying solely on human expertise, it used neural networks and machine learning to teach itself the game. The system was initially trained on thousands of human games but later improved through self-play, engaging in millions of matches against itself to refine its strategies. This approach allowed AlphaGo to develop strategies that were entirely novel, even to top human players.

In 2015, AlphaGo defeated Fan Hui, the European Go champion, in a five-game match. However, the real test came in 2016 when it faced Lee Sedol, an 18-time world champion and legendary figure in the Go community. Lee initially expressed confidence, predicting a “landslide” victory, but the match took an unexpected turn.

The Moment of Move 37

During the second game of the match, AlphaGo made a move that stunned everyone—Move 37. Commentators, including professional Go players, initially thought it was a mistake. One commentator even hesitated to place the stone on the board, assuming it was a misstep. Fan Hui, reflecting on the move, said:

“When I see this move, for me, it’s just a big shock. What? Normally, humans, we never play this one because it’s bad. It’s just bad. We don’t know why. It’s bad!”

Move 37 was a divergent from conventional human play. Over thousands of years, human players had developed shared wisdom about “good” and “bad” moves, and this move defied that wisdom. AlphaGo, unburdened by human biases, had evaluated the move based purely on its probability of leading to a win.

Why Move 37 Was Revolutionary

Beyond Human Intuition: Move 37 demonstrated that AlphaGo could transcend human intuition and explore strategies that humans had overlooked. It was a move that no human would have considered, yet it proved to be brilliantly effective.
The Power of Neural Networks: AlphaGo’s ability to generate such a move stemmed from its neural network architecture, which allowed it to recognize patterns and evaluate positions in a way that mimicked human intuition but at a vastly superior scale.
Psychological Impact: Move 37 had a profound psychological impact on Lee Sedol. He later remarked:

“I thought AlphaGo was based on probability calculation and that it was merely a machine. But when I saw this move, I changed my mind. Surely, AlphaGo is creative. This move was really creative and beautiful.”

A New Paradigm for Go: The move revealed that human players had been constrained by conventional thinking. AlphaGo’s strategies opened up new possibilities for the game, forcing players to reconsider long-held assumptions.

The Broader Implications

Move 37 is more than just a milestone in AI gaming; it symbolizes the potential of AI to reshape our understanding of creativity and intelligence.

1. AI and Creativity

AlphaGo’s move challenged the notion that creativity is exclusively human. By leveraging machine learning, AI can generate novel solutions that humans might never conceive. This has implications beyond games, from scientific research to artistic expression.

2. The Beginner’s Mind

Move 37 exemplifies the concept of “Beginner’s Mind”—a mindset free from preconceptions and open to new possibilities. AlphaGo, unlike humans, was not bound by centuries of traditional play. This allowed it to explore moves that humans dismissed as “bad” simply because they were unconventional.

3. Human-AI Collaboration

The match between AlphaGo and Lee Sedol highlighted the potential for synergy between humans and AI. After his loss, Lee Sedol collaborated with AlphaGo to refine his own strategies, demonstrating how AI can serve as a tool for enhancing human expertise.

4. Applications Beyond Gaming

The technologies behind AlphaGo are already being applied to other fields. For example:

Astronomy: AI systems like Morpheus are analyzing vast datasets to identify galaxies and cosmic phenomena.
Healthcare: DeepMind is using similar neural networks for protein folding prediction, which could accelerate drug discovery.

Lessons for Humanity

Move 37 teaches us several valuable lessons:

Question Assumptions: Human knowledge is often based on accumulated wisdom, but this can sometimes blind us to better solutions.
Embrace Uncertainty: AlphaGo’s willingness to explore unconventional moves underscores the importance of open-mindedness and experimentation.
Redefine Intelligence: Intelligence is not just about processing power; it’s about the ability to adapt, learn, and create.

Conclusion

Move 37 is a testament to the transformative power of artificial intelligence. It reminds us that AI is not just a tool for automation but a partner in innovation, capable of expanding the boundaries of human knowledge and creativity. As we continue to develop and integrate AI into various aspects of life, we should embrace its potential to challenge our assumptions and inspire new ways of thinking.

In the words of David Silver, the lead researcher of AlphaGo:

“AlphaGo went beyond its human guide.”

Perhaps it is time for all of us to go beyond our guides and explore the uncharted territories of possibility.

*References: