Why Shouldn T The Machine Be Allowed In War Games

In the film, War Games, the computer that Matthew Broderick hacks into is faced with a serious problem - whether or not to end its war games simulation. To the machine, it is simply playing a game, and there are no consequences for playing this game. However, for anyone looking to the computer for information, there seemed to be many Soviet ships on alert and ready to strike the United States. The machine in War Games had been programmed to simulate war, and come up with the best strategy for the United States were a third world war to break out. The number of people dead made no difference to the machine, and because of this, the machine did not care about how the results of the war came out. Rather, it was more interested in learning from each past simulation to better the outcome in the end. As a machine, it had no morals or values loaded into it that would tell it that any loss of life is bad, and that the best outcome isn’t thermonuclear war, but rather to just not play.
Bostrom argues that we may use the associative value accretion, or when we “begin life with some relatively simple starting preferences… together with a set of dispositions to acquire additional preferences in response to various possible experiences,” could be one way in which AI obtain values (Bostrom 189). In War Games, the machine learns that the best way to win in thermonuclear war is to not even play because it learned from playing itself at tic-tac-toe. It used its original abilities to learn a new fact about its previous function - just as Bostrom described in his