The Turing Test was created at a time when computer science was in its infancy. Compared to the computer systems that are in circulation now, Turing was using a basis of low powered computing to determine intelligent computing. The test was designed to find out if a computer can convince a human that it is responding with human-like answers.
To begin the test, there are three people who are separated in three different rooms. The main component of the test is the human judge. The human judge is there to communicate with the other two people. They will communicate with the judge using some type of text messaging system. This type of system is used to remove the auditory biases that would have occurred using the computer auditory sounds of Turning’s time. Even with the progress made with spoken computer systems such as Apple’s Siri, computers are still not able to process speech the same way as a human. The test was for the human and computer to answer questions asked by the human judge. If the judge could not tell the contestants apart, then the computer was viewed as intelligent. The reasoning behind the test is sound, but when trying to decide if the computer is as intelligent as a human, there are aspects that are missing.
An intelligent being is tough to define. There are various levels of intelligence around the world. A system designed to use language can help fracture the intelligence level of its contestants. If a computer is programmed to use a certain dialect, it could sway results of the tests in either direction. There are certain dialects that give the false perception of lower intelligence. If the human judge speaks in a southern accent or dialect and the computer was programmed in the same dialect or accent, then it would give a connection between the two. Say the other contestant speaks a northern dialect or accent or even worse English as a second language, then the human judge might read the actual human answers as having less intelligence. Using language as the device being judged, there are just too many variables at play. If a computer has reference to a dictionary, it is already at an advantage for answering vocabulary questions. If the human judge would ask a question using a word that was not familiar to the actual human, the human would be at a distinct disadvantage.
This carries over into the actual line of questioning. In Turing’s time, having a computer answer simple questions would be considered quite revolutionary. Now with the advances in technology, it would take quite an intense questioning session to fool an actual human judge. People all over the world interact with computers every day and rely on them to complete personal tasks. If a human trusts a computer to give directions using GPS all over the world, they are going to be skeptical on some of the answers in a question session. I would not be surprised at all when asking a random math question, such as 3728/42*32/3, if a computer could answer that question correctly immediately. I would expect a human to not be able to answer that question. Would that prove that a computer is more intelligent than a human?
The type of questioning that a computer would not be able to answer would be ones that involve feelings. I could ask the participants to tell of a time they were scared, why did they become scared and how they felt when they were scared. The computer answer would not be as elaborate as the human answer because they do not have feelings. They do not know what it is like to be scared. The computer would also lack the reasoning skills to tell a story. If a computer was programmed to answer these types of questions, it may come close to answering in a reasonable way. The missing parts would be ways of reasoning. Could a computer tell a complete story in a reasonable way with emotions? The emotions are what make the story real from a human perspective. The human knows what it feels like to be scared. Their heart rate will rise due to an external stimuli. The computer might know that detail but I doubt it would be able to put a story together with the stimuli and connect them in a reasonable way. Questions about feelings would ultimately be slighted against the computer and in favor of a human.
In conclusion, Turing’s test would have been revolutionary for his time. There are inconsistencies now as computers have increased memory and processing time so that they are extremely efficient in answering certain types of questions. They can answer math questions much quicker than an actual human. There have even been computers that answer simple questions better than humans in a gamelike setting such as Jeopardy. The computer just doesn’t have the reasoning skills to answer questions about feelings. These extend into human experiences. Computers do not have senses so they do not have actual experiences. These are the biases created by the Turing test.