Awry AI’s AlphaGo Origin Story: Trickery Tramples Trust in Tech

This is the cautionary tale of AI interests abusing the integrity of the epic board game of ‘Go.’

Why it Matters. AI’s origin story that AI is more intelligent than the smartest humans is a cautionary tale of publicly hyping a new AI technology as trustworthy, when the public presentation and implementation of it is actually untrustworthy trickery. It is not being transparent, forthright, honest, or fair - with the game and more importantly, with the public.

Untrustworthy-in is untrustworthy-out. Ask yourself, if Google-DeepMind’s AI technology is so trustworthy, why are its tactics to win this game so untrustworthy?

For publicity and global marketing of artificial intelligence overall, in 2016, Google-DeepMind, then tech’s top AI entity, orchestrated a public challenge and match of wits between its top AI computer program, AlphaGo, and the world’s best possible human opponent.

This rigged AI/human contest was viewed by over 200 million viewers worldwide.

1. The ‘Go’ Game. The popular game of Go is an abstract strategy board game for two players. The aim is to fence off more territory than the opponent.

This most complex board game has a 19x19 board of 361 squares which is exponentially more complex than chess’ 8x8 board of 64 squares. The game originated in China 2,500 years ago and is still played by 46 million people in 75 nations.

2. AlphaGo: The Top AI Program Challenger. AlphaGo is Google-DeepMind’s leading AI computer program that plays the board game Go. At the time of this challenge match, DeepMind was considered the leading group of the world’s best AI researchers and experts.

3. Lee Saedol: The Top Human Champion. To generate the most global media and attention, Google-DeepMind chose AlphaGo’s public opponent to be 18-time Go world champion, Lee Saedol, who at the time was the best and most revered Go player in the world. Another professional Go player said : “He was an idol and star…Everyone looked up to him.”

4. The Monetary Prize. Google-DeepMind offered Saedol a $1 million prize for beating AlphaGo in a best-of-five game match. Saedol lost the best-of-five game match 1-4.