Future of Artificial Intelligence - Essay Example

CAN MACHINES KNOW? Knowledge, according to Plato, is defined as a justified true belief. Humans are said to be the only species that possess this ability as promoted by religion, philosophy and science. However, the emergence of machines that are known to have “artificial intelligence”, which could imitate the working of human brain and even beat humans in highly intellectual game of chess, brings us to question whether machines could now know and place a claim to knowledge and learning. We will answer this problem according to the standards set by the theory of knowledge. What Constitutes Knowledge? The theory of knowledge is a branch of Western philosophy that focuses on analyzing the nature of knowledge. Different theories define knowledge and related terms like “belief” in different ways and they set conditions of varying strictness on what constitute knowledge. There are also several ways of citing what makes up knowledge, some of them are more or less “internalist” and some are also “externalist.” I would like to follow the internalist point of view which says that knowledge is a justified belief that requires the believer to be aware of the cognitive process of the given belief. If this is the case, internalism, then, requires consciousness in the justification of a belief. So if we ask whether a machine can achieve consciousness, for me, the answer is no. There are several notions in explaining the criteria of consciousness. For example, functionalism theory states that consciousness comes about when all the parts are put together in the right sort of a whole. However, I would like to believe that personal experience or ability to experience is important in determining consciousness along with the ability to think and interpret. To experience “experiencing” one needs the skill of perception and intelligence. While one may argue that a computer with sensors could function in this manner, the concept of human independence, feelings and ethics or moral capacity is missing in the whole process. Turing Test and AI “Computing Machinery and Intelligence” by A. M. Turing is one of the most controversial studies that answers the question on whether a computer can think. Here, a test is proposed to measure intelligence as well as a theoretical foundation for both classical work in AI and AI-influenced philosophical discussions of mind and psychology. The whole idea was elaborated by Paul Churchland: Put both the candidate machine and real human being in another room, out of sight and out of earshot of the judges… The judges must decide – based on a long teletyped, question-and-answer conversation with each of the two hidden communicants – which is the machine and which is the human. (p. 227) Churchland explained that the judges’ questions probe on a broad range of topics such as emotional profiles, social skills, political views and so forth. If the machine passes the test, then Turing claimed that it is conscious.1 Several machines passed this test but there were critics who remained unconvinced as to reliability of the process. S. R. Graubard is one of those who believe that the development of AI would be slower than what was widely believed and that some efforts would eventually come to an end.2 To quote: AI is something of a myth… In suggesting that a machine with vast intellectual capability was in the offing and that this new intelligence might quickly be attained, these advocates made themselves hostage to critics, who… insisted on their inadequacies. The critics resented the exaggeration implicit of those who saw only the promise of quick results and the epistemological assumptions. (cited in Crocket 1994, p. 58) Sources of Knowledge The theory of knowledge specifies the ways in which we acquire knowledge: Perception, Language, Logic and Emotion. As mentioned above, a computer or machine may have sensory capabilities and, therefore, has a partial perception capability. Also, a computer may be capable of Language and Logic but not of Emotion. Furthermore, a computer or a machine is not capable of secondary ways of knowledge, most of which are transmitted through cultural means such as traditions and beliefs. Although a computer could store information, it has no socialization skills as well as individualism to accommodate the likes of mythology, rumors, experience, and spiritual belief, among others. Although, there are those who claim that a machine can learn knowledge by itself or autonomously, there are still not enough evidences of its existence. The farthest scientists and computer technologists were able to come up are the likes of a car that could navigate its way in dessert through a preprogrammed software and a branch of machine learning called genetic programming (GP), in which software learns and evolves. These facts only show that so far machines excel only in artificial domains like chess and mathematical theorem proving. It tells us that machines are not responsive and adaptable to the requirements of solving problems that involve the physical world. Furthermore, an “intelligent machine” so far has not distinguished itself to have acquired knowledge from the artistic fronts. I believe that the idea of artificial intelligence is a misnomer and that such machines could be appropriately called “expert systems”. So far, machines have failed to acquire knowledge that can match human expertise in just those areas in which human skill is intuitive. The human experts behind the development of computers and “intelligent” machines still cannot articulate these skills; therefore the knowledge engineer cannot translate them in a set of rules and mechanical codes. At this point, I would like to state the difference between data manipulation or computational processes versus the meaning of personal experience. The former is what is behind the said “computer/machine intelligence”. A knowledge engineer or a programmer may create a program, a database or a system that could process stored information and the machine use it and analyze for future reference and use them according to its preprogrammed capabilities. An example is the computer Deep Blue, which was designed to play chess and eventually defeat human players. In a way, the data stored function as the “experience” for the machine. Personal experience, however, is different mainly because it is where personal knowledge could be drawn. It is a combination of previous activities done through various human abilities such as socialization, intelligence, reasoning, sensory perception, adaptability, beliefs, among others. Personal experience becomes important when we construct new meanings by making sense of previous meanings which are product of prior experiences. Mechanical Limitation It has been my argument that machines cannot know. As far as available evidences are concerned, there is no machine capable of learning or in possession of the ability of acquiring knowledge yet. The possibility of its development in the future is also far fetched. For a machine to know, it must have the intelligent consciousness and reason which set a human apart as an individual. It has no complete perception of the world for itself and could not translate or infuse the meaning of concepts like pain, intuition, and emotion to a mechanical simulation of the brain or specific programs giving semblance to learning. It is just a vessel or a receptacle from where data could be stored. Human intervention is still necessary for machines to function, otherwise it will remain dormant. Finally, the main argument for me on why a machine cannot “know” reflects my cultural background: A machine does not have the capacity of moral knowledge. This idea is especially important because it enables people to have true beliefs about moral matters – to believe that it is good to give to the starving and bad to torture children. A sense of what is right and what is wrong is important in acquiring knowledge and interpreting them or considering which knowledge is worth having. I think this guide our actions firmly and enable us to see more clearly how the world is. Faith in a God, the rationality of human beings, the concept of human evolution – these are important factors that makes human knowledge unique and that no machine or amount of technology could simulate. Work Cited Churchland, Paul. The Engine of Reason, the Seat of the Soul: A Philosophical Journey into the Brain. MIT Press, 1995 Crocket, Larry. The Turing Test and the Frame Problem: AI’s Mistaken Understanding of Intelligence. Intellect Books, 1994 Read More