Artificial Intelligence

Troy M. Miller

Artificial Intelligence is definitely a touchy subject for the human race. The very mention of the term conjures up images of apocalyptic societies where intelligent super-computers have either enslaved the human race or eradicated the inferior species altogether. For some, the connotation of "artificial intelligence" attacks the very core of the human spirit, the pride of our race. The very thought of an "intelligent" computer that is on par, or more likely superior, to our own brain sends chills down the spine.

Are these concerns realistic? Or are they unfounded worries of people who don't understand the issue? Some proponents of artificial intelligence insist that such concerns are the result of semantic misunderstanding. Artificial intelligence does not equate to artificial life, they claim. AI refers to only a computer that is able to "seem" intelligent by analyzing data and producing a response. One example is "smart agents" that ask you certain questions and then return recommendations based on your answers, all within a friendly user-friendly environment. Other examples include computers that can "learn" from mistakes in a limited way, such as the IBM chess program that beat Kasparov. The defining aspect of this view is that the computer is limited in it's capacity; it's intelligence is, in effect, only artificial at best.

Other supporters of AI take a more extreme view. Often referred to as a "learning computer," they describe a computer that can not only react to input, but learn from it. The computer is able to interact with its environment, make mistakes, and re-write its code to handle the resulting circumstances. This view brings the computer closer to being compared with a human; the current computer is compared to a small infant that starts out with a certain set amount of information or instruction, and then "grows" into a more intelligent, rational adult. The computer is now able to analyze, reason, and, given a situation, make an "intelligent" decision based on past experiences. Such a computer would be desirable for formulating and evaluating scenarios incredibly quickly to advise their human counterparts or, in some cases, implement the decision themselves. We see a computer of this type in the movie Wargames, where it is used to simulate military strategy.

Here, however, is where some lines begin to be crossed. To some, the learing computer has the potential to be more "intelligent" than humans - to others, it's inevitable. At the far end of the spectrum, some view the human brain as nothing more than an extremely complex computer, and see no reason why computers, when they become fast enough and big enough, can't evolve into something similar. This stance is radically different than the first view, for now the computer is much more than a tool that assists us in our daily activities. In effect, our own existence is reduced to being that of simply a complexly evolved computer. This implies that we are only useful until a faster, more efficient machine comes along, and this frightens and offends many people, giving rise to the works of fiction talked about earlier.

Where is the point, then, where we switch from futuristic probability to science fiction? The beginning of the AI spectrum, one can hardly dispute the possibility of a computer that can seem artificially intelligent in a limited way, because the early prototypes of such computers already exist. The learning computer raises more question marks, but since the basic premise is still only a computer analyzing data and making decisions, the concept is not all that unthinkable to most people. It is the last assertion that creates waves of animosity between those who claim to be pro-artificial intelligence and those who, by contrast, are labled as anti-AI.

Many people, such as myself, who are considered "anti-AI", do not deny the possibility or even plausibility of basic artificial intelligence such as described in the first two scenarios. However, the notion of a computer "evolving" to a level on par or surpassing humans is quite hard to swallow. When we begin to talk about artificial intelligence on this level, the meaning of the term changes dramatically, and the issue raises some serious objections. Basically, it comes down to the question of the differences between computers and humans. To truly be considered intelligent on the level of humans, a "being" must first be able to take in all the input surrounding it, evaluate it, and therefor make an "intelligent" decision. Obviously, a computer is quite good at doing this, but let us look at the nature of the input. Sure, a computer can handle numbers and compare results with pre-defined (or even re-defined) standards of right and wrong. The problem arises when such factors as emotions are introduced. How does one quantify happiness, anger, or distress? How does one quantify love?

The argument against this usually centers on the human itself, insisting that the ability to analyze and feel emotions is also a learned thing - an product of evolution, if you will. If humans were able to develop these capacities as a youngster, what prevents a computer from doing the same? In the end, there can be no definitive answer, because the conclusion one reaches is largely dependent on his or her view of human life in the first place. Those who subscribe to a "natural" evolution of mankind have no problem imagining a similar process with computers, aided by us along the way. Those who don't buy evolution, however, and insist that the human being possesses a spirit that is unique to him alone, cannot reconcile this belief with such a computer. In particular, man's ability to create is a trait which no other animal seems to possess. Along with genetic instinct and copy-cat behavior, humans also have not only the ability to create, but a strong desire to do so. Perhaps this is why some desire to create the most amazing thing of all - something that can also create. It is this defining characteristic which I believe makes this last version of artificial intelligence a wholly impossible one.

We see, then, that in reality the differences between supporters of artificial intelligence and those that are labeled to be in opposition are not as drastic as it may seem. At any rate, the vehement disagreements usually stem from misunderstandings of the meaning of the term, and once that is established, the argument is transferred to differences in personal beliefs about human nature, not about technical difficulties. Undoubtedly, work will continue in both areas of artificial intelligence, and most likely the basic version of AI will be realized in the near future. However, as I've stated above, I believe the quest to create an "artificial human" will only result in the realization that there are some things which man just can not do.