What the world might come to – Elysium, Los Angeles 2054, 2013 TriStar Pictures
Google’s AlphaGo has won its second game against Go master Lee Sedol, winning the first two out of a five game series. Under normal circumstances I’d shrug my shoulders and move on, because to me Go is a complicated, remote game that makes chess look exciting. But I’m paying attention because of all the stories being generated by all the technology outlets (Wired is in the lead on this) about the significance of this achievement. And I’m paying attention because it’s something Google built, and a part of the Google internal infrastructure. Like search.
I don’t believe the robot overlords are massed at the gates of our society, just ready to come crashing in to obliterate us. We have too many sociopathic politicians and leaders who are doing that already. I’m concerned because Google has created the world’s most sophisticated and powerful machine-based idiot savant. A machine that still requires human-in-the-loop to direct its capabilities to whatever end the human masters have in mind.
After watching how Google (and recently Facebook) have taken foundational computer science and turned it into a world straddling system for monitoring us for advertising purposes (and leaked it over to the NSA for a world straddling system for surveillance) I’m concerned how such systems would be used by the US as well as other countries. And don’t kid yourself. China in particular has been a top tier member of the supercomputer fraternity for decades. Other countries, such as Russia, Iran, India, and Great Britain, can purchase hardware powerful enough to run the equivalent of AlphaGo for their own secret uses. And once they do they’ll begin to use such systems to determine a strategy to run the world that will not be to 99% of the world’s benefit. Unless, of course, you’re rich and in charge already.
Whether it’s war or societal logistics, those with the best strategists and planners will, in the end, come out on top of every other group they compete with. I see AlphaGo and its descendents as the self-correcting part of any planning system, capable of refining its techniques over a very short time. Look at how AlphaGo has beaten Lee Sedol in the first two matches, and a cold chill begins to settle over you. A chilly feeling that here’s a system you might never be able to outsmart, because it not only knows what you know, but it has the ability to execute winning strategies you would have never even considered. Imagine what society might come to if such a system were used to control any protests, any dissent, and know in advance who to spy on and how to keep protests from forming by automatically identifying key leaders in any movement and controlling what kind of critical information is disseminated amongst all parties. We now have all the keys for the ultimate totalitarian state. All it takes is the right person with no conscious to come along and put them together in the right way.