In “Computing Machinery and Intelligence”, A.M Turing investigates the question of whether or not machines can think. He creates an “imitation game” in which an interrogator communicates between two separate room through writing and must figure out which is the human and which is the machine. Turing explains how the best strategy for the digital computer would be to “provide answers that would naturally be given by a man” as this would make it appear human. This immediately reminded me of the old AOL chatbot Smarterchild. For those of you unfamiliar with this screen name, Smarterchild was an online robot that you could IM and have almost human conversations with.
hilannatay: want to come over for coffee?
SmarterChild: Maybe, maybe not. You know how it is.
hilannatay: quit interrupting
SmarterChild: Why are you telling me to quit interrupting?
hilannatay: because i’m trying to tell you something important.
SmarterChild: Oh yeah? Because you’re trying to tell me something important?
hilannatay: because i love you, okay?
SmarterChild: Oh yeah? I love you too.
Turing briefly explains how digital computers can carry out these operations through programming. Digital computers are discrete-state machines, or “machines which move by sudden jumps or clicks from one quite definite state to another”. Digital computers are universal in that they can mimic any discrete-state machines if they are programmed to do so.
He then goes on to challenge his own argument by explaining the objections to it. There are nine total:
1) Theological, that God gave a thinking soul to every man and women and thus animals and machines cannot think.
2) The result of machines thinking would be so bad that we should refuse to acknowledge the idea completely.
3) Mathematics proves that there are limits to the powers of machines.
4) Machines do not have consciousness and do not really understand what they are learning.
5) “You can make machines do all the things you have mentioned but you will never be able to make one to do X.”
6) As Lady Lovelace states, machines cannot originate anything but can only do what we tell it to
7) The nervous system is too complex to compare to a discrete-state machine
8). Man is unlike machine because they do not follow a definite set of rules
9) ESP, telepathy, clairvoyance, precognition, and psychogenesis
The most interesting part of this piece, in my opinion, was when Turing claims that the best way for a machine to imitate man is to have it learn as humans do. He suggests we must program machines to simulate children’s minds and then educate them and allow them to have experiences in hopes that they will eventually learn, mature, and develops as a human mind would. He relates this process to different principals of evolution and explains how discipline must be taught through “unemotional” channels. These ideas can be connected to the 1999 film starring Robin Williams tilted Bicentennial Man. The robot acts as a maid for the family designed to simply serve humans orders. But unlike the other robots like him, he begins showing emotions, creativity, and personality. It is as if he was able to watch and learn from the family, almost in the fashion Turing describes. The family educates the robot helping him to understand basic human concepts. The robot searches for others like him, and finds one, but she is programmed to have personality and did not develop it herself as he did. The plot of Bicentennial Man gives Turing an answer to his question and creates a world and unique case where a robot can think.