Man vs. AI and the Game of Go

The ancient Chinese game of Go or Weiqi or Baduk as it is known in some parts, is believed to be the oldest board game continuously played to the present day. First originating almost 3000 years ago, it is still extremely popular in many East-Asian countries. An abstract strategy board game, Go was considered to be one of the four essential arts the cultured, aristocratic, Chinese scholar was required to master in – qínqíshūhuà – qin: a stringed instrument; qi: the strategy game of Go; shū: Chinese calligraphy; and huà: Chinese painting. Despite its relatively simple rules and straight forward pieces, Go is known to be extremely complex. The number of legal board positions in Go has been calculated to be approximately 2.1 × 10170, which is vastly greater than the number of atoms in the known, observable universe, estimated to be about 1 × 1082.

Now take a minute for that fact to sink in and try and think of all the possibilities. It is definitely more than any human can imagine right? But is it more than what technology can unravel? That was the question that DeepMind, a team of scientists, engineers and ethicists, committed to furthering research on Artificial Intelligence sought to answer. After much research and debate, Demis Hassabis, the chief executive and his team created AlphaGo, a computer program that combined an advanced search tree with deep neural networks that simultaneously deciphered the next move based on input information and predicted the winner all in real-time – essentially making it a very strong Go player that could continuously learn, make decisions and better itself. ‘We think of DeepMind as kind of an Apollo program effort for AI. Our mission is to fundamentally understand intelligence and recreate it artificially’, said Hassabis. Back in 2016, they put AlphaGo to test – first against Fan Hui, the European Champion who was defeated 5-0. The AI community had predicted that it would take at least 10 years for intelligent computers to defeat humans at Go. But here we had a computer program that had already done that. To test it further, AlphaGo was pitted against Lee Sedol, the then world’s top Go player who had won 18 world titles.

In a highly popularised nail-biting tournament, the man and the machine went head to head for 5 days over 5 deeply impactful matches that rewrote everything humans knew about Go and about Artificial Intelligence. The historic encounter and the many efforts and incidents leading up to it were documented in a riveting documentary directed by Greg Kohs that drew on the drama and adrenaline of the matches and the ebb and fall of human aspirations. AlphaGo won the tournament 4-1, but at the end of it all even the DeepMind team could not ignore the implications of the tournament or the many lessons mankind learnt from it. Watch the entire documentary here https://www.youtube.com/watch?v=WXuK6gekU1Y