weirdmindofesh

So let’s clear up one thing first. The AI isn't. It has no intrinsic knowledge of what it's doing which leads to most of the weirdness it produces. It does not understand the why. Rather it's chucking the prompts into a blender and hoping for the best. Simple prompts it's likely to get a positive result but more complicated prompts are likely to miss and miss hard. The processes are iterative, and the generator relies on the human operator to guide it. With that out of the way, lets begin. Should the users of this iterative process generator be called an artist? Yes, they should. It is ultimately up to the human to know when the process has yielded desired results that have artistic merit. An analogy could be drawn between them and a wood worker. The wood worker does not affect the processes which the wood is grown, but rather takes the result of the wood growing and shapes it. Wood workers have long since been granted the right to call themselves an artist yet, all they have done was shape the wood. Granted this is a simplification of the process but the idea is the same.

Adding on to this is something far more personal, while I was learning my own tools of the trade in digital art, I was told multiple times that it could not be art as I had never touched a pencil in the creation of a piece. By that definition a sculptor who works in clay can never produce art as they also never need to touch a pencil in the creation of their work. (Maybe they do as a tool to shape a piece.) This logical fallacy that only one type of art could be called art has never sat well with me and I chafe at the idea.

I am not suited to handle the thorny issue of copyrights; such issues are best left to those more versed in the law. My opinion in this regard is that these iterative processes are more akin to a sampling DJ. Pieces and parts are pulled from the sources and the process itself, and as I said before, put into a blender and hoping for the best. Much like the musicians the DJ sampled have little claim on the finished track, same to do the artists sampled by the program have little claim on the finished piece.

And finally, I’d like to address the claim that this program can replace a human artist all together, right now, no. IT doesn’t understand the why nor does it know when to stop iterating. Various programs and projects have shown us what happens when iterative processes are left uncheck, notably the Google Deep Sleep program which consistently generated fractal like patterns roughly around the shape of the original prompt. Going off a similar advancement in speech synthesizers which can generate near natural speech, the people that would have been directly impacted would be voice actors, and they have yet to be massively displaced by these synthesizers.

In my conclusions, I have the following: The AI does not understand the fundamental why’s of art leading to still rely on a human operator to guide it. The human operator does deserve the title artist as they have the eye to know when and how to operate the program to generate desired results with artistic merit. We should not gatekeep these artists simply because the tools they have chosen are new. Copyright is still and will forever be a thorny and complicated issue, however I believe the result of the iterative process does deserve a copyright itself. The various art AI’s will not be replacing human artists any time soon.