• Aishwarya Pothula

Computing Machinery and Intelligence

Introduction

In his paper, Computing Machinery and Intelligence, Turing puts forward the question “Can machines think?”. Turing suggests a test, famously known as the Turing test, to check for human intelligence. In the paper, he also considers and refutes myriad criticisms and puts forth a two approaches to constructing a machine that possess human intelligence.


The Turing Test

Consider a test in which a human interrogator asks questions to two subjects, X and Y, whom he cannot see to determine their gender. While X’s task is to mislead the interrogator, Y’s task is to help the interrogator determine the gender.


An example conversation in a scenario where X is male and Y is female

I : Will X please tell me the length of his or her hair?

X : My hair is shingled, and the longest strands are about nine inches long

Y : I am the woman, don’t listen to him!

Reference: https://searchenterpriseai.techtarget.com/definition/Turing-test

Now, in the proposed Turing test, if X is replaced with a digital machine, we are to see if the interrogator makes mistakes in determining which is a computer and which is a human as often as he did determining the gender while playing with two human subjects. It is understood commonly that if the machine is able to convince the interrogator that it is indeed a human, it is deemed to have passed the Turing test.


Turing himself believed that “ in about fifty years' time it will be possible, to program computers, with a storage capacity of about 109, to make them play the imitation game so well that an average interrogator will not have more than 70 per cent chance of making the right identification after five minutes of questioning”. I believe his intent in writing this paper is to normalize and promote research in building systems that can “think” without it being dismissed as utopian.


Opposing views and Refutations

  1. Theological Objection: Only man has an immortal soul. God has not given an immortal soul to animals nor machines Refutation: Don’t you think god can confer a soul into a machine if he sees fit?

  2. “Heads in the sand” Objection: Consequences of thinking machines would be dreadful Refutation: Turing believes that people who have this objection often believe that man is superior to the rest and building a machine that can think will lead them to believe that they have lost their superiority. He does not actually provide a refutation as he does not believe it to be a substantial objection.

  3. Mathematical Objection: There some questions that are either unanswerable or the answers to which cannot be verified in the system that is answering. This is a disability of machines to which humans are not subject to, making it highly impossible to build a machine which can pass the Turing test. Please refer to this explanation of the objection for a clear explanation. Refutation: “here might be men cleverer than any given machine, but then again there might be other machines cleverer again, and so on”

  4. Consciousness Argument: “Not until a machine can write a sonnet or compose a concerto because of thoughts and emotions felt, and not by the chance fall of symbols, could we agree that machine equals brain-that is, not only write it but know that it had written it. No mechanism could feel (and not merely artificially signal, an easy contrivance) pleasure at its successes, grief when its valves fuse, be warmed by flattery, be made miserable by its mistakes, be charmed by sex, be angry or depressed when it cannot get what it wants." - Professor Jefferson's Lister Oration for 1949 Refutation: “… only way by which one could be sure that machine thinks is to be the machine and to feel oneself thinking”.

  5. Arguments from Disabilities: "I grant you that you can make machines do all the things you have mentioned but you will never be able to make one to do X." X can be many such things as the ones presented in the consciousness argument. Refutation: There are no proofs provided by those objecting that a machine cannot to a certain task X. They are mostly speaking from their experience with capabilities of the then existing machines.

  6. Lady Lovelace’s Objection: “ (a machine can) do whatever we know how to order it to perform" - Lady Lovelace’s journal. This objections can be interpreted to allude that there is nothing a machine can do that can surprise (nothing new) a human because it only does what we have ordered it and it cannot think by itself. Refutation: “There is nothing new under the sun." Who can be certain that "original work" that he has done was not simply the growth of the seed planted in him by teaching, or the effect of following well-known general principles”.

  7. Argument from Continuity in Nervous System: No discrete-state machine can be intelligent as brains are of continuous-state. Refutation: Discrete-state machines can imitate continuous-state machines with a very small margin of error.

  8. Argument from Informality of Behavior: There are no set rules that a human “ought” to follow in every possible situation. On the contrary, there is a set of rules that a machine “will” follow in every possible scenario. The argument supposes that even when the world is deterministic, machines will operate deterministically, contrary to humans Refutation: World is non-deterministic, so the argument does not apply.

  9. Argument from Extra-Sensory Perception: If humans have ESP then the interrogator would exploit this to determine which of the subjects is a human. Refutation: Turing says that if telepathy is proven then the test needs to be conducted in “telepathy-proof” rooms.

How to Design a “thinking machine”

Turing feels “Instead of trying to produce a program to simulate the adult mind, why not rather try to produce one which simulates the child's? If this machine is then subjected to the education and experiences a child receives, it would obtain an adult mind. Designing a child brain would free us from thinking about the initial state of mind at birth for the adult brain, the education it had, any other experiences it faced etc, making it easier to design. Turing suggests the use of rewards and punishments in the teaching process. Another approach, which build on the previous, he suggests, is to design a system with the best sensory mechanisms that are available and then teach it to understand language.


My thoughts

What I like

  • I am blown away by the various perspectives considered, in the form of objections, right from theological to mathematical.

  • The paper works to lessen the stigma attached to pursuing such a research

  • Most of the objections considered and the refutations are still relevant today

  • With tremendous advance in compute power and storage, both approaches suggested by Turing seem possible to experiment

My reservations

  • The Turing test does not seem like an evaluation standard that can be used in the near future. The test already assumes a model that has language capability and common sense. None of the machines at present are any where near to possessing these capabilities.

  • There is no way to assess the intelligence level of agents who fail the Turing test

  • The Turing test is not reproducible.

  • Passing the Turing test maybe one of the many other things intelligent agents/humans can do. Is it a sufficient condition for intelligence ?

2 views0 comments

Recent Posts

See All