Deleted my previous posts of AI-generated art because I got my mind changed and no longer want to glorify these generators but here are how my thoughts on them evolved as someone who commissions a LOT of art:
At first I felt so liberated because drawing is the one creative thing I can’t do AT ALL, yet I have strong visions for a lot of genre-bending art ideas, which take a ton of references and effort to explain. I’ve always felt bad confusing the artists I commission. AI has been amazingly helpful in generating references, especially for the general vibes of an art piece. But I also came to the initial conclusion that it won’t replace human artists any time soon, because it’s still impossible to get the AI to generate super specific compositions. It’s especially bad with prepositions and handling multiple subjects at once. Therefore, I haven’t abandoned any of my commission plans just because AI art is available.
I realized this attitude is pretty specific to me though, because I have set character designs in mind. But what if a publisher or something just needed a generic character design to slap on a cover? I tested the limits of the AI by telling it to generate a “YA fantasy book cover by [popular cover artist]” The AI is wonky with faces and limbs, yes, but by generating variations and doing upscales over and over, I got some renders that look ready to be slapped on a book:
Some of you can probably recognize exactly which artist I put into the prompt (not naming her because I don’t want to encourage others to do this). Now what is stopping a big publisher from generating art bred from her style instead of hiring her? Nothing except $600 a year. That’s how much the corporate license for this AI costs, with “corporation” defined as having a revenue of over 1 million per year. And it’s an OPEN commercial use license. They say all paid members ($10 - $50 per month) “own” any asset they generate.
After talking to many of my artist friends, I now think these loose usage terms are pretty dangerous when artists are already systematically devalued and struggle with art theft and industries that cut corners whenever they can. People talk a lot about how AI sucks at faces and limbs, but what about non-portrait artists? I think environmental artists, background artists, concept artists, and especially artists who do surreal horror type stuff are most at risk of losing jobs to AI, even though it’s THEIR WORK that’s being fed into the AI to generate renders. There needs to be more transparency about where the images being fed into the AI are coming from, and options to limit the pool to only creative commons images or something. That or ban all commercial use, or we’re really gonna get AI generated book covers and other shit.
Right now, I think the AI user community also fosters unhealthy entitlement to art. I’ve seen people complain about the $30 tier not having enough Fast generation hours - you’re already getting hundreds of art pieces by typing a few words, and you’re COMPLAINING about a price tag that low because there’s a limit on how quickly you can get them??
These are not people who would pay a human artist the rate they deserve for their work. And honestly, I have no faith that big companies will keep paying artists fair wages when they can just buy a $600 per year license for unlimited assets to use for their million dollar projects.
In conclusion, I think AI art generators are a powerful tool and can be fun to play with, but regulations on commercial use need to come down real soon, or artist wages will decrease across industries as companies hold this hostage over artists’ heads.
