LogFAQs > #968751973

kirbymuncher posted...

Most of the people doing this I think would not be involved with creative visual art in any way were it not for the AI. So why not let them have some sort of input into the creative process and see what happens?

Because they still aren't involved: there is no creativity involved in AI art. You are not drawing or painting or sculpting anything. So-called "prompt crafting" consists of going on Reddit or one of those gallery sites and grabbing an already existing prompt for something you like and messing with random seeds until a thing you like more pops out from Stable Diffusion, NovelAI, or whatever else you might be using. Maybe it's just something cool to mess with, and that's fine---I like it too. But you see a ton of people claiming they "made" a piece and are proud of it, or just flat out claiming they are good at drawing, painting, pastels etc. and made it themselves.

You have not become creative by imagining it might be cool to see a "chibi catgirl in an autumnal forest with a wizard tower and purple sky in the background", because literally anybody who draws or paints thinks things like that. It's what you do before you start working. As you work through the forms on paper or in clay or on canvas or digital tablet, you become closer to your work and begin to express your sensibilities. It becomes your "chibi catgirl in an autumnal forest with a wizard tower and purple sky in the background", not an algorithm's "interpretation" of your prompt. Claiming an AI piece as your own work is like claiming something you paid a guy on Fiverr for is work you made.

Now, the use case you pointed out where a guy maybe wants a background because his real interest is character drawing is fine (though it would be better if he at least touched that background up a bit before using it---especially with unanswered questions about copyright). Even better, I think, is using AI generated art for inspiration or as the basis for a larger piece.

Kakapo posted...

People thought cameras and photography in general would destroy art.

Machine learning--AI--is fundamentally different than what came before it. There is active work being done on AI 2D art, AI 3D art, AI animation, AI text generation, AI chat bots, AI voice generation, and AI music generation. Some of it is already very impressive. How much need will there be for any form of human creativity in say 10 years? Companies won't pay people for things they can have a program spit out. Robots also continues to advance. In 10 or 20 years of software and hardware advancement, how many jobs will be available for regular human beings?

Those this topic is about AI art specifically, watching an unforeseen leap in a computer's ability to perform what many considered to be a uniquely human task is unsettling, and rams home the point that humans aren't special and calls into question what a future world will look like.