Artificial Intelligence

Portrait from AI

Whoa, I’m stepping into my own world of ignorance. I certainly can use the words, spell them correctly, and add the two dictionary meanings to define the field, but I’m pretty sure I don’t know a thing.

I’m now commenting on it because recently the phrase has been attached to several fields of art, where I spend a great deal of time.

The first connection is with photography. I suppose that anything powered by a battery could “loosely” be called “artificial intelligence”, but we photographers allowed it as “automatic” features. We could do it the action, but it was faster to put the camera in “A”, instead of “M”.

Big deal… we would manually select the f-stop, and the battery would power the camera to select the shutter speed.

I remember the days when the battery only powered the light meter. If the battery went dead you could still shoot pictures if you remembered the basic daylight rule.

Now, with the battery in charge of the shutter, your camera was dead as soon as the battery was dead. Within a few years the “auto” function took charge of the shutter, the f-stop, and the focusing. Of course the manual option was available, but only if the battery was charged. Hmm, that’s a clue. Manual operation needed a battery?

The last few years, especially with digital cameras, it was obvious that “artificial intelligence” was alive and well in photography, but we didn’t really call it that.

Last month I purchased a unit that attaches to my camera with a cord, then sends data to my iPad/iPhone in order to control the action of the camera. It could be called a remote controller, but it does so much more. It takes multiple shots at different settings, then blends them into a final product. It truly makes quick decisions to make a very complex photograph. It deserves the title “artificial intellence”, because it goes beyond my ability to duplicate the process.

I paid good money to get images that I could not generate with the normal skills that I have. My control of the machine is minimal, but the images are breathtaking. Question: Are they mine? Just because I paid for it, and hit the start button?

So now I jump to digital art. I’m very much beyond the arguments that define “what is art?” DuChamp settled that decades ago. Digital art just uses digital tools. If the tool is a stylus or a mouse it is still art. If the computer use a data program, it still has a programmer. Or does it?

There are programs that make fractal based designs that are spontaneous. Not common, but they exist. There was a time when “computer art” was seen as very “unartistic”, but that has disappeared for awhile now.

Naturally the argument reoccurs with AI for art programs. The typical program begins with digital images uploaded into memory, and after a few minutes an image pops out that is almost entirely created by the machine.

True the facial features may resemble the image uploaded, but the colors, the background, and the clothing may be completely different. Is it better? Is it the artist’s?

In some regards it’s just a filter. Digital artists have learned how to make filters for years, and recently even free programs make use of filters, and using them is second nature in order to get the desired results. You pick and choose the results that are revealed. You could make them yourself, but it’s easier to use a program. It’s still very custom, and tightly controlled.

This is a new AI program is not the same. If you input a portrait that is full face, straight on- it may give you a version that is a three quarter turn, because it placed markers on the original image, and calculated the turn, and filled in the data to make it right. The program is judged by the amount of correct decisions made.

The program also has gigabytes of memory for models and backgrounds of all sorts. So, the question is not whether it is art, but rather, who is the artist? In order to use the program you must be connected to a remote server to access the computing power. It takes a small fee to generate 100 different images.

I’ve spent about $50 to generate thousands of images. I’ve had a lot of fun, but I haven’t taken ownership of any of them, even if they did use my art in the initial upload.

I do upload my digital files to a company that prints the files on canvas. I sometimes send files to get images printed on metallic media. I have ownership for these files, so why not AI? And I really like the images too.

I need to ponder this…

About johndiestler

Retired community college professor of graphic design, multimedia and photography, and chair of the fine arts and media department.
This entry was posted in Commentary. Bookmark the permalink.

Leave a ReplyCancel reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.