🍬

Pros of AI Art: Incredibly easy to generate new ideas, starting points, and mood boards with a few keywords. Even easier to use as a starting point for a more fleshed out idea.

Cons of AI Art: Universally touted by the sleaziest tech bros as a total replacement for artists. There is nothing more that this crowd wants than to take what little joy there is in the act of creation and totally stamp it out. Instead of replacing manual, rote, or dangerous labor they're replacing creativity and the lived experiences of other artists because to them the end product is more important. This takes the worst impusles of the sleazy tech bro, being the Idea Man, and lets them generate content with just ideas. I use the word "Content" here on purpose as opposed to a work of art. This is the equivalent of the same tech bro showing up to drag race with their P100D tesla and out racing someone who spent years of their life and thousands of dollars building their own drag racing machine. They've "won" by skipping everything worthwhile on the journey to the end point. They claim superiority because Memes and Tech. Now they're coming for joy itself, having never experienced it for themselves

Sleazy Tech Bros delight in the idea that they'll put many artists out of work, having never valued it themselves and never having the curiosity of picking up a good pencil and doodling something. It feels like they're punishing the rest of the world for having the audacity to not dedicate every waking moment of their lives to writing software.

I've used Dalle 2 for a few months and the art it generates is decent if you squint. I've made fantasy cars, I've tried to generate furry art. I've put my own art in as a prompt and it's tried to generate variants thereof. It gets so many details wrong and is terrible at making lineart and marks in a fluid way. It's terrible at framing shots and composition. I feel incredibly guilty for typing in something like "Van Gough painting of a fox standing in a wheat field" Because neither I nor The Machine went through the life that Van Gough went through. We didn't see the sights he saw, met the people he met, or read the books he read. But the output is 85% of the same thing. It feels like a curiosity but cheapens the original artist, because now suddenly that unique life and that unique voice is now just an aesthetic and a set of keywords.

The problem is, even with art generated with a bunch of obvious errors, it's "good enough" for most people.

The widespread adoption of AI generated art means that working artists now will probably be payed to paint over or touch up what the bosses generate. There will be no more little touches and details because the AI is incapable of putting those in. Like how AI feeds off all the art posted to the internet, soon it will just feed upon itself, and like the Marvel Cinimatic Universe just be incapable of delivering anything novel, new, or exciting. Just the same content slop for the masses made by malicious and indifferent producers. It will cheapen all existing art, and any future art. I imagine "it looks like an AI made this" will be an insult at first and then a "compliment" because someone reached the same fidelity as the latest generative algorithm.

Anyway, I wish I had some positive spin on this. I don't, these are just feelings

The cons suck for sure. But, like, what can we do? It feels to me like the best we can do is try our best to influence the cultural perception of this art through conversations like these, but idk if enough people would actually care enough.

On a more bittersweet note, I feel like human-generated art is going to be not dissimilar to commenting on a cool knife or even bottle-opener that your friend has, and then being all "Wait you MADE this??" after they imply such. But then there's that fact that —unless you knew what to look for— you probably assumed without thought that they were mass-produced instead of being one-of-a-kind. So being an artist in the future would give the same vibe that modern blacksmiths or wagon-makers might have today.

And on a hopefully lighter note, I feel like creative people will always adapt. The specific trade may become rarer with time, but creatives will still find ways to put their foot in somewhere.

I suspect that there will be pockets of the internet that will have their own AI to vet whether an art is AI-generated or not before accepting it (like the AI that can detect deepfakes). Of course, this will probably create an arms race of increasingly more realistic AI art and AI that can catch the artifacts of AI, and there'll probably be false positives and negatives in such a system, but that'll be the reality of the situation.

I can also imagine some future people playing around with generating AI art when it's pretty good, but still far from perfect, growing frustrated that it cannot capture their desired subtleties before just going ahead and learning how to draw or make music or whatever themselves.