The rise of AI is nothing short of amazing. Many of those whom had either a passing interest in Chess or computers were aware of the 1997 rematch between Deep Blue and world champion GM Garry Kasparov. This match was the first time artificial intelligence was able to defeat a reigning world chess champion under tournament conditions. (Deep Blue Vs Kasparov Wiki Article). While this was seen as an impressive rise in artificial intelligence programming, it was only a mere indication of the AI to come.
Chess is a game with more possible combinations of moves than there are atoms in our universe. Before you go trying to even fathom a number that big consider that the fullsize Go board has considerably more unique possible moves in a game. It is estimated it would take all the atoms of 100 universes to equate the possible outcomes of a Go game. (Computers and Go Wiki Article)
Therefore it is not surprising that prior to 2015, the best Go AI were unable to defeat amateur ranked players. Fast forward 15 months later, AlphaGo challenged the current world's strongest Go player, Lee Se-dol to a 5 game million dollar prize match. Going into the match, Lee along with the Go community, were expecting a decisive victory in his favour. Much to everyone's surprise, Lee was only able to win game 4 after he forced AlphaGo into a questionable move, and Lee was defeated in the other four games. (Deep Learning Article)
The holy grail of AI computing still remains the general AI rather than such examples of extremely task specific AI programs. AlphaGo's implemention does however offer a peek at how more general AI might be constructed. AlphaGo utilized something akin to machine intuition in which it layered several games over top of one another to generate heat maps of successful and unsuccessful moves in given positions. This allowed it to continually play itself billions of times refining these heatmaps, searching for patterns of the best responses to common game positions.
As machine learning and the hardware that powers it continues to improve and surpass human limits in new domains, this promises to lead to very intense philosophical debates. How will governments deal with the profound change in economic structure? Will the development of AI be used to exploit individuals, or benefit them? Will we heed the warning of science fiction writers or charge willingly to a brave new world? Will AI have unintended benefits or consequences? Will humans of the future allow AI to make important decisions on their behalf? Will the exponential rate of development lead to perfected AI? What does a world with a perfected AI look like? Will a perfected AI transcend human ideals and motivations? These are the issues that we need to debate and be prepared for.