A.I. IN ART
The last two blog posts I wrote explained not only the serious medical situation I was was in, they have also pointed out my inability to currently paint. I have recently started another interior work, and I’d like to think I can finish it before Thanksgiving, but it will be some time before I can haul my equipment outdoors to do some painting. This leaves all of what I’ve done over the past summer unfinished, but eventually, I’ll get to them. So currently, instead of writign about my work, I want to touch upon an issue that doesn’t seem to go away, artificial intelligence
Throughout this year, the subject of artificial intelligence and it’s use in art has been discussed in a number of ways. It’s ability to to use creativity and style to reproduce art in deceptive ways has been the subject of a number of articles, and videos. I’ve seen panicked posts about the end of art, much in the way that I’ve seen it suggested that A. I. will be the end of writing as we know it.
My son works in the field of artificial intelligence, and he believes much of this is over blown, at least in the present. As an example, he had sent suggestions about episodes that I could write for my podcast on movie history, but they provided me with nothing I didn’t already know. In fact, I had already written on the subjects he had suggested.
The trick with A.I. is not that it’s such a deeply intelligent tool, at least not at the level it’s being currently developed, it’s that it required numerous inquiries in order for it to dig down deep enough to truly create something original. There was news about the rejection of a copyright upon a piece of art that won an award in a contest concerning mechanical use in art. The artist required many queries into the artificial intelligence programming they were using, and when I mean many, the article I read suggested well over a hundred queries in order to define the image the artist eventually painted. That’s a lot of work to narrow down an image that you would latter reproduce as a painting.
On the one hand, this probably is no different that shooting many photographs in order to find an image that an artist could reproduce on a canvas. Artists are also capable of taking those photographs and assembling them in a way that might suggest a truly create image. It can also be easy to identify paintings, based on photographs, and I hope it might become be easy to identify painted images based on concepts suggested by A.I.
But, to me, the concept still seems shaky. The art that A.I. will be able to duplicate will be not of the highest quality, nor will it be able to truly replicate an good artist’s work. At least, not yet. I’ve personally heard a number of stories considering the failure of people in developing A.I. programs, and the ability of these programs are only as good as the programmers themselves. I’m sure, that in the future, this will improve, but even then, certain styles of art will not be replicable except maybe as a computer image. After all, the value lies in cost.
While that may be important in the cyber world, in the world of art collectors, and artists, it still won’t replicate physical art unless there is a process to do the replication. So whose going to design a machine that simply draws with crayons in order to duplicate a child’s scribbles, or a young artist’s mistakes? What value is there in a machine like that except as a novelty? If a art historian can differentiate between paint from the 1700s, from paint in the 1800s, he’s certainly going to know the difference between paint applied by an artist with a brush using oil paint, from reproductive work using pigments fed through a machine and possibly sculpted by a brush mechanically.