Google's AlphaGo Victory Shows AI's Growing IQ

Google's AlphaGo Victory Shows AI's Growing IQ

For a while now, humans have been able to play all sorts of virtual strategy games against our personal computers. Whenever bored or in need of a break from that quarterly report, we might open our chess app just to see how far we can make it. This is nothing new. But, earlier this month, the tech industry reached a  milestone when an artificial intelligence program defeated one of the world's best Go players. 

First, some background. Like chess, Go is a two-person game of strategy. But whereas there are an estimated 10120 possible variations of chess games, there are 10170 possible variations of Go games, making the latter extremely complex despite its relatively simple rules. 

Go is believed to have originated in China about 3,000 years ago. Despite its age, computer programmers and AI researchers have long been aiming to develop software capable of beating the game's human players. 

Finally, on March 15, 2016, people from around the world turned their attention to Seoul, South Korea, to watch Google's AlphaGo computer program play a human world champion with 18 international Go titles. The much-anticipated matchup pitted Lee Se-dol, who has been playing for more than two decades, against Google's newly developed program. In the end, AlphaGo walked away with a 4-1 victory and a $1 million prize, while the tech community celebrated what has been called overcoming "the last remaining great hurdle" in the race to create game-playing artificial intelligence. (Recall when IBM's Deep Blue supercomputer outperformed world chess champion Garry Kasparov in 1997 and when IBM's Watson system beat "Jeopardy" champion Ken Jennings, who had a 74-episode win streak on the TV quiz show.

In a post-match press conference, Lee speculated as to what led to the 2016 AI victory, saying, "It remained unfazed psychologically and stayed focus. In that regard, I don't think humans can beat it, even though I hesitate to admit that AlphaGo is above humans in Go skills yet." Lee added the loss made him realize he needs to rethink his game and commit to more studying – a little ironic given the fact he has pretty much two decades' more experience than his inanimate opponent, which was developed in 2014. 

So why is this relevant to eCommerce, you ask? Though the match indicated what might be limits for the human Go players of the world, it confirmed the next stage of advancement for the human AI developers of Silicon Valley. Amazon, Google, IBM and Microsoft are all engaged in a "platform war" to control the future of artificial intelligence and, in turn, what is likely the future of technology. In fact, the experts at IDC predict that by 2020 – just four years from now – there will be a $40 million market for machine learning applications, 60% of which will run on the software of these four tech giants. 

The outcome of the AlphaGo-Lee match was a much-needed win for technology. However, this is not to say there aren't still hurdles to be overcome. Just last week, Microsoft unveiled Tay, a bot created to understand the art of conversation. Tay is able to automatically converse with Twitter users by mimicking the latter's style of speech. Needless to say, the chatbot soon began spewing racist, misogynist rhetoric, and Microsoft put Tay's account on pause less than 24 hours after launching it. 

Artificial intelligence still has a ways to go, but the tech industry is evidently up for the challenge. From Apple's personal assistant Siri to Amazon's Echo feature Alexa, AI is popping up all around us. Even if you're unfamiliar with Go, it is matchups like the one on March 15 that can sow us how powerful this type of technology is becoming. 

Mar 30th 2016 San Diego Media

Recent Posts