Saturday, August 2, 2008

What Can't Be Seen...

One of the enduring mysteries of science is the person sitting three rows away on the train. Seriously. As much as we know about the form and function of a human body, how close are we to an artificial creation that is the equivalent of one? Honestly enough, I’m not all that sure; I don’t keep up that well with current scientific developments. I’m pretty sure that the biggest hurdle to such a creation, however, is the mind.

As a video gamer, I’m fairly familiar with the progress we’ve made in that department, at least as far as artificial intelligence is concerned. Since video games use a computer to provide a challenge to the player, they have to be able to make tactical decisions in a lot of different situations. Advance Wars is one obvious example; it’s essentially a war game. The computer has to be able to evaluate the situation, attack and defend intelligently, and so on.

The progress that AIs have made in that respect is obvious. The first Advance Wars AIs were, quite simply, pathetic. They made tactical decisions on basic priority lists, to the point where a specific unit that had no tactical value could draw fire for more valuable units. As time passed and more games were released, the computer became much better at providing a challenge.

In fact, there were many maps in the first Advance Wars game that were blatantly lopsided. The player was at a major disadvantage to the computer, and if there was a human player in control of the computer’s position, victory was assured. Those maps either don’t exist anymore or have been more balanced, because the computer has become more capable of providing resistance.

Even if the AI in Advance Wars could mount a challenge that equaled that of a human, though; even if the AI could play the game better than any human, it still would not be the equal of one. After all, that computer can play Advance Wars, yes, but I can put Advance Wars down and play Pokémon Diamond. Or MegaMan Star Force 2. Or… well, you get the idea. That is the challenge that faces modern science: how does one create a computer that can match a human in all aspects?

Frankly, I’m not even sure if it’s possible. Again, I could be way off here; I’m not keeping track of current developments as much as I should be. But I haven’t seen a computer yet that can exceed its programming. The computers in today’s world are programmed to perform a task. They are often far more capable at that one task than a human would be, but cannot do anything outside that task. The computer that controls a robot can be programmed to vocalize a greeting, such as, “Hello!” when it identifies a human in front of it. But when I see someone I don’t know, I can say any number of things, from “Hello!” to “Good afternoon!” to “What’s your name?”

Often, I don’t even know which one I’ll use until right before I use it. It’s not like I go walking around thinking, “Okay, I’ll walk up to that person and say hi.” Which is another aspect of a human intelligence that I’m unsure about replicating: spontaneity. I can decide to do something for no apparent reason. Very little factored into the decisions I made about what to do next as I rode down the highway. (I’m coming back from vacation as I write this. Yay laptops!) I just felt like playing games for a little while, and then I took out my iPod… none of that was determined by reasons that a computer would be able to find. I just felt like it.

As I said in the second part of The Growth of High Technology, an essential part of being human is human emotion. How does one code for that? Sometimes, when I’m faced with a serious challenge in a game, I quote Captain Sisko, among many others (“Fortune favors the bold”) and just charge in. Sometimes, humans act outside of self-interest or in unpredictable ways. Whether motivated by altruism, love, or just sheer insanity, humans don’t always act logically. Yet computers are driven by logic; how does one create computer code that can replicate such quirks of humanity?

There are several predictions flying around as to the power of computers in the future. Certainly, they’ll grow ever more powerful; it’s not like they haven’t been in the past. I’m not sure, though, if we’ll ever design one that can imitate a human perfectly. There are just too many facets of human behavior that can’t be easily measured or copied. Regardless of one’s beliefs about a soul or other aspect of humanity beyond that of scientific confirmation, there are too many things about humans that can’t be seen.

No comments: