By Dylan Walko,
Tech & Innovation Writer
Self-functioning AI technology is not just an innovation of the future any longer, it has been around for years now. First seen on a global stage with IBM’s Watson participating in Jeopardy, these machines have the ability calculate algorithms and process data in a matter of seconds. This was seen once again over the past few days in South Korea as Lee Sedol took on the AI robot AlphaGo in the ancient game of Go.
Go is similar to that of chess and requires the same analysis and projection that chess does, thinking 10 moves in advance and contemplating the results of every decision. AlphaGo is a Google product that was created in London at the Google Deep Mind lab. The two men who created the intellectual inception of AlphaGo are Demis Hassabis and David Silver.
The initial results of the competition were as anyone would have expected, AlphaGo handed Sedol three consecutive losses in a best-out-of-five campaign. Wired stated that Sedol was emotionally distraught after the third loss and apologized “for not being able to satisfy a lot of people’s expectations.” Yet in a true Cinderella story fashion, Sedol was able to force AlphaGo to resign in the fourth game. After this the tide has turned for Sedol as they enter into the fifth and final game of the series.
The rational so far behind the blunder that transpired for AlphaGo has been the color of the stones they use to play. Black traditionally always takes the first turn followed by white, in the fourth match Sedol was given the white stones, going second allows the player to get a head start in points and Sedol felt that with this slight advantage the playing field was now even.
The key as well was masterful play by Sedol and a lack of comprehension for AlphaGo, during one move the machine was cornered on the board and in turn used a subpar counter move. From there Sedol was able to capitalize on its poor play. Eight moves later AlphaGo finally recognized and recalculated the board to find its error but by that time Sedol had virtually put the game away.
On a larger scale, this marks a milestone in AI unlike any other before. As stated by The Verge if Google’s Deep Mind team can create an AI machine that has the ability to evaluate and execute strategy at a world-class level in a board game, there is truly no telling what it can do for the world at large. By creating smarter processing and analysis it is possible to use machine intelligence to generate greater efficiency.
This type of AI may seem much narrower in theory but it does reach a greater underlying concept. Many AI platforms at this point have the issue of only containing a set amount of pre-programmed templates. It does not adapt to new environments or situations that it does not understand. Things such as AlphaGo and Watson are machines created to “think on their feet” as a human would. Creating new solutions to issues on the fly, instead of going back to the basics that are installed within it.
AlphaGo is much more than a simple machine that beat one of the world’s greatest Go players, it speaks to the idea that artificial intelligence is constantly growing and is gaining its own sense of self. It has the ability to think and process obstacles just as we do, maybe just better.
A version of this article appeared in the Tuesday, March 22nd print edition.
Contact Dylan at