Can a Machine Know?

What do humans mean when they say they know something? In the common scheme of things, they mean that they have learned some facts about a particular subject. In addition, they mean that they are satisfied that they have an understanding of the sources, relationships, and significance of that subject. If that is all knowing means, then machines could also have the capacity to know. However, knowledge, a human capability, is based on learned facts, experiences, judgments, growth, change, interactions, and other active processes. The concept of knowing requires the description, consideration, and application of a multifaceted process lacking rigid rules, goals, and boundaries. None of these attributes are available to a machine, so a machine's capabilities would not constitute the requirements for a machine to know.

If knowledge is fact based, then a machine can know. Because machines have the power to reproduce material stored in their memory, it could be said that they know the facts and are capable of using them. All existing facts could be converted into digital form and imputed into a machine in any language. Since knowledge can change based on acquisition of new facts, more facts could be continuously entered so the claim that knowledge may be factual can be satisfied.

Humans use information from different areas of knowledge to make a reasonable guess that can be proven over time. For example, the existence of black holes is unproven but humans assume they exist from looking at images of distant stars having materials extracted from them by apparently large, super-gravitational forces. Machines, however, are unable to investigate ideas due to their limited reasoning abilities. They cannot speculate that the existence of a black hole may have an indirect proof. If knowledge were more than just purely factual statements, then a machine would fail to know.

Can a machine differentiate between right and wrong with its limited understanding of human nature? In a court case, a machine could determine whether or not a person has broken a law based on strict application of a statute and the pure facts of the case. If the accused has committed horrible deeds, a program in a machine's memory bank could recognize the things the person has done, deeming him guilty or not guilty of the accusation. However, criminal cases are seldom black and white because of possible mitigating circumstances in which a person, who perfectly fits a machine's classification of a murderer is able to evade jail time due to special reasons. One could premeditate murder or kill out of self-defense, but would a machine know that? To a machine, any type of killing would be considered as murder or slaughter. Based on the dictionary's definition, ethics are moral principles that control or influence a person's behavior. A machine would not be capable of determining the ethical aspects of such a situation.

A woman who was beaten by her husband for twenty years stabbed him to death when he was passed out drunk. The judge gave her lifelong probation but no jail time. The judge knew that it was immoral to jail this woman, who was suffering more than she deserved. He would have known how to include the current state and directions of the social and political environments in an application of the law. The judge overcame limitations in the law, which a machine could not, by exploring possible uncertainties and eliminating bias. A machine would be unable to comprehend the morals of the situation to come to a reasonable conclusion while the judge, on the other hand, would use reason, a way of knowing, in coming to his judgment.

Possessing knowledge enables one to be aware of the opposite side of a situation or argument. Since knowledge must be applied to a large range of environments and situations, not just simple calculations and factual statements, a machine's ability to answer questions with pure facts is not considered knowledge. To determine if a person is innocent or guilty, a machine would refer to the definitions imputed into their memory chips and match it with the person's data to come up with an authentic reply. However, a machine would be unable to offer an opinion on anything. It could only analyze the facts and provide a one-sided answer. A machine has no conscience or the ability to opinionate; it would fail to create knowledge because of its lack of reason.

Humans use their subjective perceptions when they judge something. They have an idea of what something could be or could not be. A machine can be equipped with sensory devices that would enable it to perceive what humans call a sheet of paper. It could scan it and form a picture, measure its dimensions, and take its weight. The machine could then refer to a rule in its memory which delineates what a sheet of paper should be. But is the machine looking for a piece of paper? Does it have a plan for it, or know what to do with it like humans would?

The machine has senses that indicate what something is, but all it has done is perceive. According to Plato, a famous Greek philosopher, a human would put more emphasis on the idea of the form of an object, not on defining what the object is. Humans have a general idea about any particular object, and that object would be an example of that form. The object could be annihilated, but the idea still remains. This is where a machine would fail to have knowledge. In the absence of an object, it would no longer perceive it. The existence of an object then does not depend on the perception of it. The human mind, conjectures Plato, does not rely only on perception to know that something exists; our own minds know what an object is without relying on perception. Even if a machine were given rules to guide it in defining something, the machine could