The Turing Test refers to a proposal made by Turing in the year 1950, it examined whether machines have the ability to think. Turing discussing the question whether machines can think which in itself seems meaningless, but using a party game then known as “The Imitation Game” Turing proves that the questions deserve some discussion(Oppy and Dowe,2016). As early as 1950, Turing was convinced that in due time there will exist digital computers in the future that will perform better in the imitation game (Oppy, 2016).
Turing addresses a possibility that there is a continuous state of machines(unchangeable output) completely different from the discrete state machine(memory based output) (Oppy, 2016). His claim is that even if humans are continuous state machines there exists a possibility that discrete state machine will be able to imitate human behaviors sufficiently especially in the Imitation Game. Turing himself claims that even if machines can imitate human behaviors there are limitations on the notion of a machine. For instance, he says that we should allow the possibility that engineers could construct a machine that works but its manner of operation is not satisfactorily described by the constructor since they applied experimental methods. Turing knew very well there would be a class of machines which to him are potentially much larger compared to discrete state machines. As a result, he was confident if there existed properly engineered discrete state machines they will accomplish in the Imitation Game (Oppy, 2016).
Due to continued criticism and controversy about the thinking machine, Turing states that we would be able to recognize an intellectual machine through the type of responses it expresses. We can interpret the machines utterances as proof for grief, pleasure, anger, misery, and warmth among others. To be able to come up with this kind of machine will require fitting a machine with affective states, and sensors similar to making an artificial person. This is similar to the claim that an algorithm of the human consciousness can be developed and fitted in a machine to perform the same functions as a human (Boltuc, 2008). Interactions of the real world are not just a string of words but through tones and action that precede it that we usually refer to interpret behaviors. Hence, a stronger test for human-like intelligence might include speech and image processing or an algorithm introspection of human mind fitted into a machine (robots) to give it feelings and speech or a state of mind.
Chinese room
The Chinese room is a thought experiment coined by Searle who asks the audience to imagine themselves as monolingual English speakers who are locked in a room and offered a Chinese script and writing (Searle, 1980). Then you are presented with a third batch containing formal symbols with more instructions in English which mean you can now be able to correlate the script and the writing using the symbols. Looking at the scenario we can refer to symbols or the sequence as a question and completion or correlation of the writings and script as the answer. The man will become skillful of matching the symbols to a point no one can distinguish him with a Chinese person. However, the fact remains that he is not Chinese and does not understand Chinese, let alone thinking in Chinese.
According to Searle (1980), your reaction to this situation is similar to a computer “Script Applier Mechanism” a story understanding program. Searle believes that since the computer lacks the knowledge of the Chinese language just as he does both of them do not know anything. Contrary to strong AI, despite how intelligent a computer may behave or the programming it has since the symbols it has are meaningless, the computer is then not intelligent. Searle’s argument was countering Turing’s claim that machines can one day be able to act as humans or human consciousness can be replicated into the machine. According to the Chinese room, a programmed computer would only know what it has been programmed to know and nothing more. For this reason, it is recommended in the future tests such as Turing and Chinese room should not follow a script or instruction. This would allow AI to create algorithms and links that are fitting to certain situations.
Arguments Against Replicating Consciousness
Several individuals have indicated that replicating consciousness’ is the only way we can be able to explain the easy problem as well as the hard problem of consciousness. The hard problem is essentially asking why a human has subjective experience and are phenomenally conscious (Chalmer, 1995). Human beings are said to have subjective experiences which mean there is something that it is like to be them. For instance, the means through which someone understands how color is formed and work but could only see black and white. Since currently there exist no results reporting the existence of consciousness, it would be a milestone for surgeons to isolate the visual cortex area representing color while preserving activation patterns. A new theory should be invented reason being, if we understand the mechanism and function of the hard problem clearly we can make an algorithm to implement into the machine (Boultuc, 2008). We are only capable of explaining the easy problem which involves simulating consciousness but the artificial intelligence created would just have a human-like cognitive function, not consciousness. Boutltuc (2008) using three premises explain why we possibly explain consciousness with science and why he believes that strong AI can be made.
The counter-argument to this theory is that it is impossible to tie algorithm to a human mechanism (Dournaee, 2010). According to Dournaee, the first solution is to believe that the hard problem is not executable. Basically, it is very hard for us to determine when we felt and experienced something (2010). Secondly, Dournaee (2010) state that is for the humans to believe the hard problem is not representational. This is an indication or claim that human cognitive often mismatch with each other making it even harder for artificial brain to replicate it.
Dournaee (2010) explain that for AI to have consciousness it must be able to know when it is conscious. I do not believe the knowing of realization of consciousness matter in both human and AI. An algorithm can be predetermined that fit the AI, and it still can have consciousness. An approach to not needing the realization moment of consciousness is reportability. Reportability plays an important role in consciousness, without reportability, it is impossible to tell whether or not one is truly conscious, An example of this is when glancing at a photo and asked to report what was in it. The participants will make some mistakes and falsely report on what was there. When telling what to look for the were able to see the image fully, therefore, becoming conscious. Similarly, when the first AI algorithm is created it will not be as perfect as a human, but through mistakes and fixes to the links to the situation, it will experience consciousness.
Another argument for replicating consciousness is by using the study of neural mechanisms of phenomenal judgments that can be applied to solve the hard problem. Under this argument, consciousness is stated to be physical and can be produced with a physical mechanism (Argonov, 2014). A pertaining mind-body problem is there are two basic alternatives in philosophy: first, it is possible to study consciousness and control as an objective matter fraction. Secondly, consciousness contains immaterial degrees of freedom that can never be controlled nor observed via objective techniques. Argonov believes that tests can be made to check the existence of materialism (representing all the theories that regard human beings as being a combination of the physical body as well as immaterial soul). All this is really hypothetical because Argonov believes that science is not advanced enough test intelligence robot. But when the time comes, this could be used for it. He states that the scientific key to the mind-body problem is human being’s ability to talk about consciousness or in other words produce phenomenal judgments (Argonov, 2014).
Counter to this argument is the“qualia” which is used as an example of problematic consciousness property that is used as an argument against materialism. It is impossible to describe the qualitative properties of human perception as well as imagination verbally. According to Jackson (1982), we are unable to explain the essence of a red color to a person identified to be color blind. Similarly, we cannot tell for sure how another person including a normal color vision perceives the color red. It could be possible that person is calling it red but subjectively does perceive it as blue.
I agree that we will never know what one person view and experience the world is like, and it is different from person to person. But we can all agree on that we know the color that represents the objects. Let say someone sees the color red but other see the color blue. Let us presume that red is truly “red”. The person might see blue, but the blue for them is still “red” for us. It could be that their perception is blue but they called it red. We cannot say that their representation of “red” is different that our.
Turing via his proposal The Turing Test stated that there will come a time that digital computers or intelligent computers will be in existence and they will be a replica of human consciousness (Oppy, 2016). He indicated that digital computers will evolve as intelligent machines that will, more importantly, play the Imitating Game better. Turing believed that a time will come that artificial intelligence will exist with thinking computers that will be playing the Imitating Game better. According to his argument, we can only judge a computer's intelligence same way we judge students understanding or skills by asking rapid and unexpected questions. He believed that humans can at a time be capable of having conversations with the intelligent computers for the purpose of solving the hard problem of consciousness. For instance, if the machine is capable of answering the questions asked correctly then we can refer to the machine as intelligent (Oppy, 2016).
Some authors claim the Turing Test to be chauvinistic since it only recognizes intelligence in the things that has the ability to sustain conversations with human beings. The Turing’s 1950 proposal paper has concentrated much on digital computers failing to foresee other entities such as analogue computers, aliens, and animals that could pass the test. Authors who critique the test ask whether Turing believed that only digital computers could become thinking machines (Oppy, 2016). Some authors indicate that by 2000 the time Turing was certain a digital computer with thinking capacity would exist passed and until now AI on machines has not been achieved.
I agree with this statement, but the creation of AI is still young and currently, we do not have the technology to do more testing pass the conversation aspect. With some early testing with autonomous conversation and empathy in AI, it shows some promises in the future with creating a full algorithm that can replicate human consciousness.
Conclusion
It is clear that without ontologies or conceptualizations underlying knowledge there are no chances of having the vocabulary for representing knowledge. According to Type-A consciousness, there is no epistemic gap between physical and phenomenal truths, therefore, our knowledge and surrounding experiences are therefore based on one’s past experience. There are two types of Type-A consciousness including eliminativism and logical behaviorism. Eliminativism is the view that consciousness does not occur and therefore, there exists no hard problem. On the other hand, logical behaviorism claims that consciousness exists and mental states are a representation of the actual human behaviors. Reliable sources indicate that it is possible to replicate human consciousness and install it in machines as programs. This is only possible by first understanding human consciousness or the state of mind and the behaviors associated with it. After fully understanding human consciousness we can develop an algorithm replicating human behaviors and transfer it to a machine making it an intelligence machine as foreseen by Turing. This is possible as according asynchrony of consciousness believes that consciousness is created by interactions between two parts of the brain i.e. the geographically and functionally parts. With the knowledge, it is easy to isolate brain functions and replicate as Artificial Intelligence on robots or other machines. All in all, AI is still a new field that was not focused on since last decade. But with the booming in technology advancement, we can expect the first AI with the algorithm that resembles a human to the basic level of interaction and consciousness in the near future.