Can a Machine Know - Essay Example

Can a machine know The question of whether "a machine can know" may become one of the central and most important ones in the Twenty-First Century. The question raises many different problems. First, what does the word "know" mean Second, how will human beings recognize when a machine, most probably a computer, is starting to 'know' Third, what kind of knowledge does a machine display already The word know needs to be defined. According to Webster's dictionary the word know has the following meanings: 1. perceive, as a fact or truth 2. have information of; be acquainted with. 3. recognize or distinguish in comparison (Websters, 1995) Do any machines already show some of these meanings It is clear that many machines "have information of" many different things. Thus a machine as simple as a toaster "has information" regarding how hot it needs to be and how much time is needed for various kinds of toast. It has sensors that cut the heat off and push the toast out. At the most complex level, computers can store massive amounts of information and can provide instant access to it. So at this very basic level of "knowing" many machines already do know. Turning to the third definition of knowing, do any machines "recognize or distinguish in comparison". Again, the answer is yes. A car can recognize a whole host of complex facts, from the road conditions, to the air temperature to its own state of repair. It distinguishes a full gas tank from an empty fuel tank and acts upon the information through providing a warning light for the driver. The car distinguishes a door being closed from a door being open, and will beep when a door has been left ajar. The car, or rather the computer within its engine, is distinguishing between one situation and another. So the third definition seems to have been met. But perhaps the most important question is whether a machine does or ever will be able to "perceive". Essentially this is a questions that deals with whether a machine will ever be able to "think". Will we turn on a computer one day and have it say "hullo, I am here", and know that it is saying it. Thus the basic question is whether a machine will ever be alive. The idea of a thinking computer, with all the benefits and risks involved, has existed virtually since they were invented. From the robots of the 1950's science-fiction to Star Trek computers gone out of control, to Hal and on to Data in Star Trek: The Next Generation there has been a pervasive fascination and/or fear with the thinking computer. The fascination involves the apparently limitless potential of a machine that combines the raw processing power of a computer with a human ability to reason. The fears stem from the same possibilities: will human beings stay the dominant life-form with a thinking computer However, as the slowness of the development of robotic technology has shown, the creation of Artificial Intelligence (AI) is much more difficult than previously thought. Computers can learn from their mistakes, can teach one another and even repair one another, but none so far has shown any sign of self-awareness. They do not, or at least do not seem to, think. At the moment the processing power of a computer is similar to that found within simple animals such as an ant or a flea. Such animals do not have "brains" in the sense of higher animals such as mammals. They have neural centers where their relatively simple life functions are organized and implemented. Ants probably are not self-aware in the manner of a human being, and yet they are "alive". The question arises of whether a computer will show self-awareness in a mammalian or perhaps even human sense when, within a few decades, it has a similar processing speed to the human brain. Here this analysis moves into perhaps highly difficult definitions of "life", "thought" and perhaps even "spirit". As far as biology is concerned the human brain is made up of a network of neurons between which various connections are made. Two neurons, or long lines of neurons, either have a connection or do not have a connection. They are either on, or off. In terms of number, they could be seen as either a 0 or a 1. This is exactly the way that a computer works: even the most complex programming works through a series of 1s and 0s. The similarity between the computer's way of storing information and the human's is remarkably similar. But there is there a sense of self, of the "I" which each of us wakes up with every morning, that is specific to biological life At the moment it must be stated that human beings simply do not know. Computers have yet to achieve a processing power that equals that of higher animals, let alone that of the human brain. Inevitably, however, they will do. In the "Turing Test", created by one of the fathers of computer science, Alan Turing and imaginary situation is set up. A human being has two conversations. One is with a person, the other is with a computer. If the human being cannot tell which is which then the computer is said to have passed the test. The problem arises as to what "passing the test" implies. Can a computer be said to be "thinking" or "knowing" if it passes the Turing Test The answer is obviously no. A sufficiently fast machine with a huge memory could be programmed with a large enough number of human questions and human responses to deliver a human answer to almost every question. It would also be able to give a vague random answer to the few questions not in its memory. This would simulate human response in a purely mechanical way. The computer would be no more thinking in this case than the toaster knows it is burning bread to make toast. The computer, as the toaster, is performing purely mechanically. If a computer can mimic human speech perfectly, and perhaps even look and act like a human being with advances in robotics, the question moves from whether it is knowing, or thinking, to whether it actually matters. Is a perfect copy of human behavior any different from actual human behavior This is a profoundly disturbing question. As many philosophers have stated, and perhaps most famously Descartes with "I think therefore I am", the only thing that any individual can be sure of is their own identity and existence. The rest of the universe, including other human beings and their supposed thoughts, might be just an illusion. In such a world, the difference between a computer mimicking human ways of thinking and knowing and actual humans is not important. To conclude, at the most basic level, machines do "know" things. At the most important level, at the level of considering whether machines "know" to the level at which they show self-awareness we may safely argue that they do not. At least, not yet. If computers do start to show signs of self-awareness as they reach human processing speeds, then a whole series of questions regarding machine rights and what life is will be raised. _________________________________________ Works Cited Webster's New American Dictionary, Signet: 1995 Read More