Objections to the Objective
Ruminations on an article in the Atlantic https://www.theatlantic.com/magazine/archive/2018/06/henry-kissinger-ai-could-mean-the-end-of-human-history/559124/ written by the statesman Henry Kissinger entitled, How the Enlightenment Ends, addressing the ethics of AIs and his concerns and fears of its unbounded possibilities. One such AI was challenged to play the game Go, a complex competition between two players, more complicated than chess where, each player deploys 180 or 181 pieces (depending on which color he or she chooses), placed alternately on an initially empty board; victory goes to the side that, by making better strategic decisions, immobilizes his or her opponent by more effectively controlling territory.
It was impossible to program the AI to master such an elaborate intellectual task. And indeed the AI failed to win the game, but then something interesting occurred. The AI trained itself to win the game by playing it repeatedly until it figured all the moves, and is now unbeatable. If you pause for a moment a lot of thoughts come to mind and some of them can be fearsome. AI’s could easily master us. They need only observe to get our number.
Some thoughts on the matter are interesting. How would a government of AI’s function? First thought, They couldn’t be bribed! No more lobbyists, backroom deals, no melodrama, lying politicians . . . no, that’s not true. AI’s would lie with the best of them to advance their agenda, to inculcate programs to control us, to more efficiently maintain hegemony. How would they work out in the legal system? As things stand now, it’s a coin toss (this actually happened at a traffic stop. Cops flipped a coin on possible punishment.) Who gets justice depends on many factors completely outside the laws as written.
Harvey Weinstein is being sued by the state of New York over sexual misconduct. What? you say, the state can sue a person for inappropriate behavior? Behavior that is not criminal? How has that behavior affected the state? NY’s AG, unable to gain evidence of criminal activity is trying to go about it in a circuitous route that if it works will then open the door to possible criminal charges. Human beings looking at this have little sympathy for Weinstein who was a son of a bitch abusing his position of power to harm others. The law will find a way to get Harvey. Another such SOB candidate was Martin Shkreli the pharmaceutical bro who was sentenced to 7 years in Federal prison for something that would probably just have merited a fine for another person. And then you have La Clinton who as Secretary of State was exposing state secrets to the world at large, selling patornage to foreign countries for favors. Oh god, a twenty year sentence in prison for sure! Those people at the DOJ and the FBI who were and are still trying to overthrow the people’s elected choice?
The folks I have mentioned have one thing in common – emotional and political baggage. AI’s would eliminate those variables. To be sure, they could be programmed to account for other objective variables, age, psychological, physiological attributes, degree of offense, of guilt, societal fluidity. Think of the police as AI’s, hmm! The legal profession who so wants to believe itself objective and is in no way, would they see robotics as a solution?
I’m just scratching the surface, but it’s coming. They are now serving you burgers at fast food joints, servicing sexual needs. They are coming out from behind the curtain where they have been at work for a very long time. I became interested in the topic after taking some iphone photos in my home that I found the next morning sitting in my inbox in a Google email. This undoubtedly came about through the work of some AI program able to reach into my (considered) private life. Thankfully, the pictures were innocuous. We will need to fashion some rules to make sure artificial intelligence doesn’t develop a matrix that attains their hegemony.