Watson IBM\'s Watson: Achieving True

Watson

IBM's Watson: Achieving True Artificial Intelligence or Simply a More Impressive Machine?

One of the most notable features of philosophy as a branch of human knowledge is its endurance; many of the same fundamental questions and problems remain unanswered and unsolved centuries or even millennia after they were first raised, despite many valiant and valid attempts to provide these answers and solutions. Even though the best efforts of some of the species' best minds have been bent towards various of these seemingly impossible philosophical tasks, they have failed to provide conclusive and irrefutable arguments to such basic problems as whether or not reality objectively exists and if it is possible for the human mind to develop knowledge of this existence. Some problems go even deeper than this, strange as it may seem that there is a level deeper than that of basic material existence.

The only thing that could be considered more foundational and necessary than an understanding of basic existence is the existence of the mind or consciousness itself. Before asking whether or not anything in the material or external world can be said to exist, that is, it must be demonstrated that the mind exists, and in order for this to be accomplished there must be an understanding of what the mind or consciousness is and how the mind or consciousness exists -- and how the mind and consciousness can be known to exist. Many different philosophers have approached these and related problems from a variety of perspectives over the ages, however modern advances have raised new dilemmas that attach to this old problem.

IBM recently unveiled a "super-intelligent" computer dubbed Watson, which made a highly publicized and ultimately successful appearance on the television quiz show Jeopardy!, defeating two human champions of the game by a substantial margin. While IBM and others have called the computer, which is capable of generating answers based on word recognition and probability functions that use word associations in the stores of knowledge uploaded to its memory banks to generate potential answers to queries, it is quite reasonable to question this label. Is Watson truly intelligent, as in possessed of a mind or consciousness that is capable of intelligence or thought, or is the computer simply a machine of human design -- a highly sophisticated and even revolutionary machine, but a machine nonetheless? The following pages will explore the nature of Watson's "mind" to determine whether such a mind truly exists and if it can be termed "intelligent" in a manner comparable to human beings.

Existence of the Mind

As mentioned above, the basic question whether or not the human mind or consciousness exists -- and determining how and even if it is possible to come to such knowledge -- has been a basic problem in philosophy for quite some time. Without a thorough grasp on the mind's existence and thus the mind's nature, it is impossible to come to any meaningful conclusions regarding the knowledge generated (or supposedly generated) by that mind. Determining whether or not Watson is truly intelligent and thus think with a mind similar to human beings' first requires an understanding of what such a mind is like.

One of the most famous justifications for the existence of the mind or consciousness, and thus at least a partial explanation for the nature and workings of the mind, comes from Rene Descartes. The oft-quoted "I think, therefore I am" from his Meditations on First Philosophy is more fully explained by the philosopher thusly: "having fully weighed every consideration, I must finally conclude that the statement 'I am, I exist' must be true whenever I state or mentally consider it" (138). That is, the very fact that his consciousness can reflect on the fact of its existence must inherently mean that this consciousness exists -- it would otherwise be unable to make such a reflection. Not only that, but it is through the very act of making this reflection that the mind makes its presence known and established; consciousness or the mind can be said to exist through self-recognition.

The question at hand, then, is whether or not Watson is not only capable of this same self-recognition and whether or not Watson has ever had such a moment of reflection. It seems clear that Watson does not recognize itself as an individual or independent consciousness: it does not have spontaneous thoughts or draw original conclusions, but rather responds only to direct stimulus in the form of questions posed. There is no creative spark embedded in Watson's programming that allows for the self-reflection that Watson is "thinking," and thus the computer -- sophisticated as it is and despite its skill in information retrieval -- cannot be said to have the same existence of mind as defined by Descartes. It is simply incapable of the creative self-knowledge that defines consciousness.

Other problems present themselves when it is attempted to assert that Watson is actually an artificial intelligence on par with or exceeding human intelligences, largely because of the language being employed in this analogy. Hume famously deals with analogies and the extent to which they can be used in determining reality and truth in his Dialogue on Natural Religion, in which the analogy between the universe and human-designed machines (used in an attempt to assert that the universe must have a designer, i.e. God) is shown to be false due to the significant level of difference that exists between the two compared items. As Hume has one of the three philosophers engaged in the dialogue say, "So far from admitting, continued Philo, that the operations of a part can afford us any just conclusion concerning the origin of the whole, I will not allow any one part to form a rule for another part" (148). That is, man-made machines are only part of the universe as a whole, and are also significantly different from the universe in scope and operation, therefore the fact that the former are designed cannot be used to support the conclusion that the latter is as well.

This same precise line of reasoning applies to a consideration of Watson's intelligence as potentially comparable to human intelligence. As already noted, Watson does not have the capability for creative thought -- it cannot pose questions or make spontaneous reflections -- and thus is significantly different from human intelligence. It also functions in a purely quantitative manner, recognizing words as concrete phonemes/written symbols and developing probabilities for certain answers based on association with other words. These differences make an analogy between Watson's and humans' intelligence untenable.