Art and artifact. (computer graphics programs) (Column)
by Robert Bixby

I recently finished putting the final touches on our November feature on fine art, written by fine artist Lee Noel Jr., formerly of COMPUTE.

The material he provided was exciting and arresting--art that could be appreciated by anyone. In fact, you probably have had in your possession a work by one of the artists; Joni Carter's work has appeared on postage stamps.

A couple of the artists were involved in creating the software that creates their artwork. Another was working on the hardware level, stringing together machines, sensors, and output devices to create something that would result in an experience for the viewer.

At the same time I was working on the feature, I was reading a pile of science fiction books--Joe Haldeman and Larry Niven--culled from the local used-book emporium. So it was inevitable that I began to think about pushing the limits. When Isaac-Asimov created Andrew the wood-carving robot in his classic science fiction story, there was no question in the minds of people who saw Andrew's carvings whether he was creating art. But Andrew itself was only a tool produced on an assembly line. Only because of a defective positronic brain was Andrew creative.

Many people who program--perhaps most--eventually sit down to create a graphics program. I wrote a few and enjoyed the process, and here's why. The interest in text and data files lies in their meaning, and there are only a few things you can do to words with a text editor and still have words that make sense. By contrast, you can do almost anything to a graphic, and it can still be visually interesting.

In my efforts, I created something I called a wallpaper processor. It would rotate an image 90 degrees and superimpose it on the original image, move the image to the right a set number of pixels and then repeat the process. Depending on the original image, the result would look like the very busy wallpaper favored in the early part of this century, with intricate patterns repeating every inch or so.

I won't make a case that my wallpaper processor created art (although I managed to publish some of it in literary magazines), but if it were art, would the art be my art? Or could a case be made that the computer created the art and my only contribution was a signature?

What if, instead of giving the computer a set of fairly complex but rigid instructions, I had informed the computer about aesthetics--showing it how to achieve balance withouts symmetry, to use a variety of shading techniques to provide an interesting set of textures? I might even have designed an expert system that mimics the cretive processes in an artist's mind.

Alan Turing, the English computer visionary who was a member of the team that breached the codes of the German Enigma machine during World War II, devised a test to determine whether a computer was capable of thought. His test involved having a person interact via teletype with either a computer or another human being (a teletype was the only input device they had at the time). If the human operator was unable to tell whether the interaction was with a computer or a human being, then you could say the computer was capable of thought and was, in a sense, human.

It seems to follow that a computer that can create original art indistinguishable from human art--even human art created on a computer--is an artist and human in this way.

I don't think Alan Turing considered what to make of the programmer who created the program that was capable of thought. As a creator of something indistinguishable from humanity, is the programmer elevated above the huamn level? As frightening as these things are to think about, they are close at hand. It's easy to tell van Gogh from his brush, but it's more difficult to distinguish the creative programer from the capabilities of the program code.