Can a Machine Know?

Can a machine know?

Whether it is reciting the news, ferrying documents between offices or delivering food in hospitals, there is little the latest robots cannot do. Robots are automatically operated machines which replace human effort . But can they "know"? My stand is that though machines cannot do everything a human can, they can perform certain functions and they do "know" in this respect. A machine cannot fall in love or win a beauty contest, but then again, humans cannot do everything a machine can.

When does someone, or something, since the situation calls for this, "know" something? There are many definitions of knowledge, some of which limit it to humans only. The essence of most of the definitions, though, is that Ð''knowledge is the confident understanding of a subject, potentially with the ability to use it for a specific purpose '. How does one express understanding? We can know someone has understood something only when they perform accordingly or exhibit their understanding in gross methods. Given this, machines are able to perform most acts to perfection, far more effectively than humans, I daresay. Then, it can be said that machines understand, since they perform, and they "know" since understanding correlates to gaining of knowledge. I do not claim that machines understand the issue, but only the instructions; therefore, they do not comprehend the implications and reasoning behind their actions, but they "know" what they have to do.

Robots perform their functions intelligently, not just repetitively, or by remote control. They determine what is required, perform the task, and monitor performance, changing their actions as needed in the real world . Their detection of the real world is similar to us. Their sensors can be physical detection, like when a switch is bumped, or radiation detection, which includes visible light. How different is this from the way humans function? We pick up signals, like the energy radiated from other humans, analyze the situation, determine what is required and perform accordingly. I hardly see a difference in the way robots and humans act.

A common argument to differentiate machines from humans is that robots cannot act on free will. Though some can make choices, these choices are programmed. Machines operate on "fuzzy logic" whereby, there are a fixed number of choices among which they may pick and act on. Humans, however, can think outside the box, and always the concept of Ð''exceptions' is present for us, absent in machines. But are we not programmed as well, by our social background, education, exposure, etc? Though humans can absorb knowledge they choose to and reject knowledge they are not tuned in to or interested in and machines cannot, whether an individual can "know" or not does not depend on whether he is capable of making choices. I do not know archaeology, but I know biology. If I "know" then a machine which knows how to mix chemicals but not how to run a race "knows" as well. It is true that though I do not know archaeology, I can choose out of my own free will to know it. On the other hand, a machine knows only what its programmer decides it should know. However, this does not change the fact that the machine has knowledge, though only that which pertains to its area of use. The argument is not that machines can choose what to know but that they can "know"; however limited their knowledge may be.

Machines cannot feel emotions, have desires or make value judgements as they do not possess ethics or morals. They cannot know through introspection or intuition, or be inspired to push their potential. Therefore their ways of knowing are very different from ours. Machines gain knowledge through other ways; for example, memory. Memory is a very important part of retaining knowledge, and computers and robots have much more reliable memory, than humans, which will last till the programmer decides to unload that particular programme. Their memory is used to repeat actions, remember people, etc.

Some machines have the ability to voluntarily link previous information stored in their memory with new information and analyse it like we do using both inductive and deductive logic like us humans . However, machines cannot detect fallacies. For example, if I fed a computer with the data "All pigs are pink" followed by "Mary is a pig" it would generate the information "Mary is pink". From this, it is obvious that machines do not have common sense or practical knowledge like humans do.

AIBO (Artificial Intelligence roBOt), a battery-operated robotic dog from Sony, behaves like a living animal, learning and growing through experience . AIBO understands voice commands, and it simulates feelings with sound and action. It can recognise visual face patterns and owner's voice.

. Like AIBO, some robots have the capacity to learn through acquaintance and experience, which is how they come to recognise their owners. Perception also plays a role in the knowledge-gaining process of machines. I might perceive my friend's mood from her actions and act accordingly. A robot, on the other hand, may perceive that its owner's glass is empty and refill it. The things a human and a machine perceive are very different in quantity and quality. Humans' perceptions also differ from person to person, since we, too, pay attention to different things to different extents, based on how observant each of us is and what we are interested in. Some robots can respond to verbal and social cues. Robotics pioneer Rodney Brooks and his fellow researchers have created a robot which, when praised, responds with a smile and when criticized, drops its eyes in

...