Turing Test commonly refers to Allan Turing’s proposal in 1950 to answer the thinking ability of machines. In an effort to answer the problem of whether machines have a mind or thought, Turing suggested an “Imitation Game,” where a machine’s ability to think is determined by how much it can convince an interrogator that it can respond to questions in a manner indistinguishable from humans. However, in general terms, Turing Test is a behavioral test for thought, mind, or intelligence in entities with minds. This essay focuses on arguing about the reliability of the Turing Test as a measure of the presence of mind or thought in machines. Turing Test is a reliable test for measuring the presence of mind or thought because conscious experience is unique to each entity since a machine can be considered not to have a mind based on the subjective conscious experience of humans.
According to the test, for a machine to be viewed as thinking or having a mind, then it should be able to deceive an interrogator that it can think like a human (“Dualism (Stanford Encyclopedia of Philosophy)”). Thus, to understand the reliability of the test to prove the presence of mind or thought in the machine, it is crucial to understand the meaning of the mind. In response to the body-mind problem, dualists assert that both physical and mental states are real and distinct. Mental processes and states are essentially phenomenological processes and states. When applied to Turing Test, dualism asserts that the right conscious experience must mediate appropriate output to the given input. Since thoughts are conscious experiences, and machines cannot have consciousness; hence machines cannot think. It is important to note that dualism heavily relies on the presence of consciousness to describe an entity as having a mind or thought. However, basing the presence of the mind of a machine on a subjective conscious experience cannot be used to evaluate the reliability of the Turing Test.
The question of consciousness of a machine has been one of the major arguments against the reliability of the Turing Test to test thought. For machines to have thoughts, they should be conscious to feel emotions and other sensory responses as humans. Chinese Room argument by John Searle is one of the major arguments against the reliability of the Turing Test as a measure of the presence of mind. Searle argued that regardless of the human-like behavior of a computer granted to it by a program, the program does not give the machine understanding, mind, or consciousness. In his argument, Searle presented a thought experiment known as the Chinese Room Thought Experiment, based on the premise of the successful construction of a computer that behaves like it comprehends Chinese. The machine uses Chinese characters as input and offers other Chinese characters as the output by following instructions given by a program. Searle says that if the machine passes the Turing Test by working effectively, a human Chinese speaker can be persuaded that the program is a live Chinese speaker (Searle, 3). Thus, for all questions asked by the person, the computer gives appropriate responses convincing a human Chinese speaker that they are talking to another human who speaks Chinese. Searle asks whether the machine is simulating its understanding ability or understand Chinese literally.
Searle then introduces a scenario where he manually runs the program to pass the Turing Test. He says that if he is closed in a room with enough stationery and an English version of the computer program, he produces Chinese characters output by processing input of Chinese characters according to the program’s instruction and passing the test. Thus there is no difference between the roles of the machine and him in the experiment (Searle, 3). But since he does not speak Chinese and, in extension not understand the conversation, so does the computer. He argues that the machine cannot think since it does not understand; hence it does not have a mind. Thus, the test cannot be used to test for the mind in the machine reliably.
However, arguing that a machine does not think based on a subjective feature does not make the Turing Test a reliable measure of whether the machine thinks or has a mind. A significant omission by the consciousness argument to prove the presence of a mind is the conscious thing, something to be felt. Each organism has conscious mental states that are specific and distinct for an organism if at all consciousness existed. Thus, consciousness is not limited to humans only, but it is widespread, and since conscious experience is subjective, then each organism can have conscious mental states, which is unique to it. Due to the connection of each subjective experience to a single point of view, it is not feasible to determine objectivity in any conscious experience.
In the same way, we cannot identify and determine the conscious experience of a machine because we cannot understand the conscious experience of the machine, just like we cannot for an elephant or a rat. Thus, consciousness should not be a factor in measuring the presence of thought or mind in a machine. If thinking cannot be a measured based human definition of consciousness, a machine that passes the Turing Test has thought or mind. Moreover, rejecting the reliability of the Turing Test, there would be no means of knowing beliefs in other minds.
Works Cited
“Dualism (Stanford Encyclopedia of Philosophy).” Plato.Stanford.Edu, 2021, https://plato.stanford.edu/entries/dualism/. Accessed 22 Apr 2021.
Searle, J. R. (1980). Minds, brains, and programs. The Turing Test: Verbal Behaviour as the Hallmark of Intelligence, 201-224.
