On The Turing Test

The Imitation Game or Turing Test has become the hallmark of the question - ‘Can machines think?’. Turing predicts that by the end of 20th century, machine would be able to think. Humans have been on moon after that and machines are still at the dawn of thinking. Thinking. The paper is philosophical one and ‘think’ is a vague term which cannot be determined in simple yes or no. Turing has therefore, given a method, a test through which we can assess if machines can think. Simply put, the Turing Test states - if a machine is able to fool a human into believing that it’s a human, then we could say that the machines can think.

Turing proposed this question and test when computers were only in labs. The idea had some discussion around it even at that time. Turing refers and answers to around 9 arguments on why machine cannot think. The most compelling argument is by Lady Lovelace -

[machines] can never do anything really new

This argument has it’s merit. Even after 150 years (Lady Lovelace’s paper was published in 1842), machine still cannot do anything new. We still have to train our models. New is just perhaps another word for creative. Can computers be creative, can they write a sonnet for example is unanswered in the paper. So can machines think? I would like to substitute another word for ‘think’. Can machines take decisions? The answer is a straightforward Yes. Not simple if-condition-decisions but more complex ones - is this a picture of a cat or a human.

There is a lot of difference between fooling a person into believing that it can think and actually thinking. A machine would be able to tell the difference between good and bad based on what it’s taught, but can it answer this question on it’s own? Can it think on it’s own? A human would be able to answer this question quite easily. The answer might be different for everyone but the dullest of men would be able to do it.

Coming back to decisions, machines can take decisions. They can decide what the next word in a sentence would be, in that sense they can think/decide what to say next but can they invent new words all together? This again goes in the realm of creativity.

The last section captures a very interesting question - we are what we are because of what we were subjected to. If we create a child machine, and subject it with the education we might get an adult brain to which we are directly comparing machines. Turing has mentioned education and ‘other experience’ to be be given this child machine. We will have to engineer this whole process. We were far from it back then, we are far from it today also.

Turing test is easy to decide if machines can or cannot think, but is it sufficient? I don’t think so. In Turing’s own words:

“Can machines think?” I believe to be too meaningless to deserve discussion.

The answer to ‘Can we say that machine would be able to think?’ - Turing predicted that perhaps by the end of the 20th century we will be able to answer. The century has passed and another score and we still don’t know. Perhaps by the end of this century we will have the answer.

Updated: