I’m not interested in developing a powerful brain. All I’m after is just a mediocre brain, something like the President of the American Telephone and Telegraph Company
In the 1950s, Alan Turing, a prominent English Mathematician and Computer Scientist, asked himself a basic, yet intriguing question, “Could a machine think like we humans do?”
Building on this idea, Turing wondered that if a computer’s response were indistinguishable from a human, then the computer could be considered a thinking machine. This motivated him to lay the foundations of the famous “Turing Test”.
The objective of this article is to explain the Turing’s Test and enlist its objectives and fundamentals and how Alan Turing derived the famous “Imitation Game”. So, in order to test the intelligence of computer machinery, Turing devised a game consisting of 3 Players-A, Z and X. The Player A is the interrogator, who is communicating with X and Y through an interface. Out of X and Z, one is a machine, and the other is a human being, and both try to convince A that they are human beings, through this “chatting session”. The duration of this session wasn’t defined by Turing. So, in this session, if A is able to detect and distinguish successfully between the two, as to which one is a computer and which one is actually a human being, then the machine isn’t intelligent enough and is declared to fail the test. However, if the person A is unable to do so, then the test is carried by taking different interrogators, that is, different A, and the results from the different tests conducted by changing the interrogators are recorded. If the computer prevents from “failing” a particular percentage of the test cases, it is declared “passed” for the overall test. Again, here also Turing didn’t define the exact pass-percentage, and hence usually scientists have assumed it to be 66% percentage, or one could say a 2/3 factor. That is if the machine passes in 2 out of every 3 test cases, it is deemed intelligent enough.
Alan Turing defined some ground rules to authenticate the results of a test. For example, if a supposed “Turing test” omits the human and uses machine for both the
players (X and Z), then it is not a “Turing test”. Similarly, if a supposed “Turing test” has the actual-human counterpart trying to give computer-ish answers to fool A, then also it is not a “Turing test”. Further, if a supposed “Turing test” deceives A by trying to hide the fact that a Turing test is being conducted, and informs A that he is chatting with 2 humans, and concludes that a lack of suspicion on A’s part is sufficient to claim that A could not distinguish between a human can computer, then also it is disqualified to be a “Turing test”.
The first computer program which passed Turing’s Test successfully was “Eugene”. The program claimed to be a 13-year-old boy from Ukraine – two factors that could be used to excuse any grammatical errors in the computer’s replies as well as its ignorance of more specialized forms of knowledge. Only 33 per cent of the judges ‘Eugene’ spoke to had to be convinced he was a human and the conversation was only five minutes long with each.
For this reason, and a few others, a lot of computer scientists have stopped viewing the Turing Test as a credible way to assess artificial intelligence. However, that doesn’t mean it’s completely useless. As it stands, the Turing test passed by ‘Eugene’ isn’t insignificant, it speaks a lot about the sophistication of the new robots, and the main take away from this is perhaps that the goalposts have moved. Turing’s original test came from a time when computers the size of buildings were less powerful than our current smartphones – the fact that it’s no longer as relevant as we first thought is a thing to be celebrated as well.