2021-22 SotA Anthology 2021-22 | Page 152

Could a Machine Think ? KIERAN PERRY

Could a Machine Think ? KIERAN PERRY

In this essay , I am going to consider the question of whether machines can think , and argue that this is not possible . I will first define intelligence within machines , referencing ideas of Alan Turing . Using this I will outline why I believe that intelligence is not enough to prove that a machine can think . I will consider the objection that machines are improving exponentially and thus may achieve consciousness and therefore thought in the future , but I will discuss why I believe this is unlikely to occur . I will use the Integrated information Theory to support this viewpoint , leading me to conclude that it is not possible for a machine to think .
To begin with , to be able to determine whether a machine can think , it is necessary to define what thinking is and what it involves . One line of thought is that thinking arises from ( artificial ) intelligence , and it could be argued that many machines are already at this point . Machines such as Deep Blue certainly display a level of artificial intelligence as they are able to beat a human player at a particular activity , in this case chess ( Bory , 2019 , p . 632 ). It may be stated that , because these machines are intelligent , they must be able to think , but I believe this needs further examination .
Alan Turing supported the idea that intelligence and thinking were interconnected , as he devised a test that he believed would show whether a machine could think . Turing believed that , if the interrogator could not discern human from machine , then the machine was intelligent , and therefore could
think ( Turing , 1950 , p . 442 ). This view suggests that intelligence is therefore all that is required for thinking .
I would argue that the Turing test is certainly a test of intelligence , and that many machines in the modern era would be able to prove that they are intelligent by fooling an interrogator . The reason for this is that machines have merely been programmed to give a certain response based on a number of different inputs ( Turing , 1950 , p . 440 ). Meaning that even if they have a very large number of possible answers to the questions that may be given , at some point they will run out of responses .
This leads me to believe that the only way a machine or any being could truly think is if it is conscious . The reason for this is that consciousness allows us to take in information and genuinely contemplate it in order to choose a response that we believe will produce an outcome that fits with our desires ( Duignan and Rey , 2009 ). This differs from the machines that display intelligence because they are not responding in particular ways out of a desire to do so , but merely because they have to ; it is what they are programmed to do . Furthermore , conscious thought is influenced by a number of factors such as our mood , whereas a machine is not affected by this . Consequently , I believe that machines cannot truly think , as consciousness is necessary to meet the requirements of free and independent thought .
Because it seems likely that machines will always be programmed in such a way so that they are
SCHOOL OF THE ARTS 2021 / 2022
152