I was seriously expecting, a year ago, to finish my trial of their services and make it an even-more-premium Pro account at this point; but times change fast, don’t they? It was certainly worth it a few years ago.
It was fun while it lasted, sure. I got to watch a number of interesting Trojan Horse was a Unicorn presentations on animation, I learned from some top artists for some Hollywood favorites, I even got to hear about the process which many other artists went through in going from concept to product.
However, it’s been slowly degrading over the past year, and Art Station’s refusal to make an enforceable statement on “AI Art” and its presence on the platform was, if I’m honest, the only real nail the coffin even needed.
It’s not like I’ve made a lot of sales on it; realistically I’m not sure I ever made even one. That wasn’t why I was there; I was looking for an outlet for art assets like Sounds.com has been for audio. I was looking for a number of video courses on art in practice. Most of all, it used to be that I could find and start a conversation with just about any major creator, which seemed too good to be true.
Now, every single post is either part of the protest against AI Art on the platform—which I am, meagerly but proudly, a part of—or something which may very well be AI generated itself. True work is something a trained eye, with a mind knowledgeable about the profession, can see; particularly in contrast to other elements submitted to the platform.
I won’t try to tell people what their own vision of AI Art should be, but you should critically listen to both sides of the argument. The difficulty in producing a portrait, is what we admire in it. Punching letters into a prompt and making your GPU suffer to produce a passable image is not art, it does nothing for you; and it runs on the stolen work of others.
To produce such art, programs like DALL-E require pre-training systems like CLIP (Contrastive Language-Image Pre-training). This means that they need to consider a wide database of material from established artists, and the prompts associated with them; and then produce a new item matching features from that compressed database by comparing the description with the associated prompts.
Is there a use for that? Absolutely. It could do wonders for natural language processing alone, which has a hand in many wonderful tools. However, saturating an established site for human artists with material prompted by a user, with potentially no artistic background, and using it to compromise the living of the very people who brought us to this point isn’t just unseemly, it’s artless in itself. It’s losing track of the very reason the site exists.
AI use is not AI “art”, art is an act of personal creation. Feeding a prompt into a text box and receiving dozens of interpretations in a matter of minutes is not an act of creation. We have to bleed and grind through all of our bad lines and improper shading before we can ever create something truly spectacular, and the sheer reality of that feeling is what drives me to do this.
So, in summary, am I willing to pay a premium for a site which promotes artwork made for pennies on the dollar, if that? Absolutely not. It doesn’t benefit me, and it doesn’t benefit my audience in any way.
Here to stay? Probably. Make no mistake, though. This is the Visual Arts equivalent of Napster all over again. If you spend hours doodling a drawing and pouring your effort and your soul into it, you should be financially compensated for that.