So here we are, at the third and final part of this article series, where I take a stab at the question of what the rise of AI means for Artists. Are we going to be obsolete soon? Should we start applying for jobs at McDonalds? Should we start working for exposure like we’re brand new again? The answer is no. Stop the doom and gloom you dingbats!
Grab a coffee and settle in, this is a long one. View it in a browser or in the app!
Before we get into it, I would like you to take a moment and muse on a few questions. There are no correct answers; what you believe in your own mind is all that matters.
What makes art, art?
When you look at art, what is the primary thing that attracts you to a piece?
Are there common themes, feelings, or technical aspects you seek across art?
Have you changed mediums or technical approaches in your own work before?
Do you make art to pay bills, or do you make art because it is part of who you are?
If you collect art, is it for yourself, or primarily an investment asset?
AI good AI bad AI
I want to start by expressing some personal thoughts about what creative folks have been doing with AI, prompts, and why the anti-AI movement is insanely misguided. I think folks are exhibiting knee-jerk reactions to problems, before solving similar ones that have been around long before this new technology.
First off, good prompt writers are fucking poets, if not straight up witches. I think that learning to manipulate prompts properly for output is practically analogous with magic, alchemy, and how spell binding works in traditional pagan practices. A pinch of this seed here, an utterance of positive intention here, a negative prompt there, now stir it up with n iterations, and let’s see what we get. At what point does magic become science? That is not what this article is about, but I will probably get into this super esoteric good stuff in another article soon.
I understand why artists are worried about their art being hijacked into training models without their knowledge. The shit that went down recently with Epic Games being fuckwads, is a great example of why people are frustrated. It begs the question though: are these same folks equally zealous about their (copy)rights when it comes to the blatant negation of any Creative Commons licenses once you agree to share your media on most social media networks? Dig into that clusterfuck here if you like. This is bad stuff.
My point isn’t that these folks are wrong; on the contrary, I feel their frustration. I have had my work literally stolen and used to promote the Gathering of the Juggalos, among countless other random events over the years. ICP gets sued by tons of large companies for other copyright infractions annually apparently. Maybe unwanted/bad press is good press?
As AI grows as a toolset, and regulations grow with it, it would be nice to see some protections, but I wouldn’t hold my breath. Nobody has done shit to protect us so far. Please see my last article in this series about what copyright actually does for us as non-corporate individuals. We have a very complex new world we have to adapt to, with the fact of our work being fed to AIs is a part of that. Go write your congressperson or something if you really think it will do some good. Good idea?
Here’s a big mind-fuck: nobody actually has a concrete answer on whether you can copyright an AI image right now. Does “human author” apply to the prompt that the human author wrote, or to the generated output from the prompt? Well fuck.
So instead of posting pointless Anti-AI art that has essentially turned into a meme at this point - effectively negating the “movement’s” point - maybe start looking at whether you want to start using AI in your creative workflow or not. Both are good. Here’s why:
AI Wanna Use It!
Great, do it, do you.
AI Don’t Wanna Use It!
Great, don’t, do you.
Glad we got that all out of the way so easily. On a less sardonic note, I had a discussion with Robek recently, where we hit this topic a bunch. He said something colloquially about generative multi-K PFP projects, but I find it applies entirely to the AI discussion. The question is: What is your intention with your art?
This is the only question that matters in my opinion, especially if you make art because it is a part of who you are. Perhaps this is a case of the ends justifying the means, as long as nobody is getting hurt against their will along the way.
Next up, I want to talk about different “types” of art, but want to preface it by clarifying that I’m not talking about music at all here, because I simply don’t know enough about it, and also because I feel it warrants an entirely different, dedicated discussion.
Traditional Artists
I just don’t understand how AI art can ever be a threat to the livelihood of traditional artists. Let’s be clear here, I’m talking about physical art - painting, illustration, sculpture, live installation and performance, you name it. AI has a long way to go to outpace any of these. Here’s a free idea for those that are way more talented than I am with a brush or pen: replicate some AI generated images with oils or acrylics, or even pen & pencil. Call it Meta AF but make the F in AF look like two lines attached to an I. Thank me later by giving me a % of your sales.
Sure, the increase in digital art may have led to a decrease in traditional arts over the last few decades, but think about how many new artists have been able to discover their creativity without the costs associated with going to art school, buying supplies, and even figuring out how to transport/ship/handle final pieces. There was adoption, and there was adaptation. If anything, this increased the value of physical art. Many trad artists also work with software. Do what works best for you, and never forget that physical art will forever adorn the walls of homes, offices, museums, and even on the streets.
Digital Artists
I understand how uncomfortable it is to think about an AI generating stuff for an agency, or your friends no longer getting custom pet portraits from you as AI seems to do it pretty good, pretty fast, and pretty cheap (mostly free). Here’s a tough pill to swallow that is probably going to lose me some favor here: you need to change what you’re doing to make income from art if it’s that easy to replicate/replace your art with some crap generated in Lensa or Wombo.
Photoshop and Procreate didn’t put painters out of business, and while AI is much more complicated than that, it isn’t about to kill the design industry. In fact, you’re already using AI in Photoshop for a number of things, in case you didn’t already know. Just roll with it, adapt, and use the tools to your advantage.
Now if you’re talking about your absolute need to create art, because if you don’t it’s going to start manifesting in really fucked up ways in your life, well, you have arrived at Casa Orb. The only thing in this entire article I can say with near absolute certainty, is that if you think people will stop valuing your work that you truly believe in, as AI starts creating comparable visual and audio equivalents… well your concern is fucking irrelevant. So long as you are clearly focused on your intentions, whether it is to communicate an idea, or just to let your body and mind become a conduit for creation, you’re going to be okay. Create value, don’t create based on your perception of value.
I mean don’t forget to get a job and pay rent and stuff till you make it, but I’m telling you, AI is probably not coming to replace you. The financial focus on non-creative sectors of business outweigh the returns that would be had by replacing artists in creative industries. All industries suffer from poor leadership that undermine ethical values, sure; the hard fact is that cutting corners yields higher gains in corporate models, and capitalism rewards boss-people that boost profit margins.
To clarify my point, the money that could be spent replacing you with highly trained AIs in order to sell art for profit, would be miniscule compared to the returns of investing that very same money in industries like healthcare, finance, automotive, etc. Guess where the money folks are likelier to go upon realizing this?
This brings us back to the topic of art valuation, which I touched on in the first article. Yes, I am one day very probably going to enjoy gesturing at my NFT showcase monitors on my walls displaying dope AI art and say things like “Look at my beautiful Tormential pieces,” the same way I would point at physical art on my wall and say, “Look at my beautiful Jaclyn Atkinson piece.” There is nothing wrong with this, stop telling people what they can and can’t like you fucking troglodytes.
And again, as markets get oversaturated with AI art, the same way digital art did for a while after tools like iPad + Procreate became easily available, we will likely see scarcity become relevant again. Not in quantity, just in medium; if there are 5K AI images of a diamond for every 1 that’s not AI generated, guess which diamond is going to shine more brightly?
3D Artists
In the context of more advanced software, such as Maya, C4D, Houdini, Blender, etc - the more complex the AI (see: part II), the larger the degree of difficulty in training an AI properly; and the more complex the software UI and its components are, the more challenging it will be to train AIs to comprehend them properly. A lot of the restrictions for algorithms here will likely be that the programs we use for CG manipulation are designed around mouse/keyboard inputs, and geared towards human comprehension of 3D space.
The point I’m ungracefully trying to push here is, fuck teaching an AI how to use a human tool, why wouldn’t you teach it to just create its own engine with which to generate the results we’re talking about? As an example, Eyes of Lamia’s concerns are absolutely valid in my opinion, but I feel that AI isn’t exactly “thinking” about things the way we are. In fact, notions of lying, confidence, etc are all irrelevant when it has no inherent, or experiential moral human compass, or even a sense of self. Back to the Turing test we go.
So let’s go back to this point of it building its own software/interface vs teaching it to use existing programs. You’re going into chatGPT and saying stuff like “copy a mesh,” and “directional normal in relation to a spatial location,” all in [insert software name]. Earlier, I said folks who are masters of AI prompts are basically poets. Well in the case of prompting an AI to use/direct usage of existing software, imagine just asking it to create a mesh, or explore spatial directionality without context. OK, now think about further extracting the terminology and going back towards the thing we actually want to create on a base level. “Generate a three dimensional dildo.” That’s a clear prompt!
The prompt for the above image was based on Eyes of Lamia’s original query in chatGPT, entered into Stable Diffusion’s txt2img: “a mesh along evenly scattered points in a circle in Houdini, facing outwards based on normals facing away from the circle center”
Is it going to default of its own accord to doing it in some 3D software? No. Parties funding AI tools to work in 3D are already focused on building AI-centric engines, and subsequent interfaces to direct/prompt the AI to create objects. This article came out about the evolution of AI 3D point cloud generators a couple days ago. Now it’s important to note that while it mentions Google’s DreamFusion, it avoids mention of the fact that while Point-E outperforms it, the computing resources necessary to push quality to a point of discernibility from human-made CG are unrealistic. Well for now anyway. Here we go, circling back to the complexity/training problem again.
So in my humble opinion, it makes zero sense that 3D artists will get replaced anytime soon, though there is the potential eventuality that all of the software companies like Autodesk make will eventually get replaced with something else, or just integrated with AI tools. It’s hard to think they would stop investing in what they’ve developed and cold-turkey their users after decades of investment; there’s got to be a transition to keep it lucrative. For example, Adobe’s Photoshop, after having integrated with NFT and AI tools, continues to provide the exact same tools we have learned to use over 35 years. Shit, NVidia; the leading GPU company on the market, came out with GET3D that generates models that are absolute rudimentary trash at best. Sheesh.
We just don’t know which way regulation will go in the next few years (see: part II), and it makes sense that workers’ rights will kick in, along with copyright and trademark laws. I feel that Autodesk would be pretty pissed to learn that rogue AIs are being used to output monetized educational content that openly uses their software.
A Rough Conclusion
I want to be clear about my perspective here; I’m not neutral. Though I am not worried about true artists being replaced in any way, I am concerned that we are in a very rough transition period. I think that bleeding edge technologies like blockchains and AIs can have symbiotic relationships wherein the chains maintain immutable ledgers of the origins of an image, and subsequent iterations, derivations, and even tributes to it. Who knows what else is out there right now.
Historically, things go pear shaped before any functional regulations are devised, and implemented. We’re still trying to figure out whether these algorithms we’re training understand what the full purpose of a dildo is - I mean most humans don’t understand what pleasure is for shit’s sake. Is this a race to get the AIs to understand us, or for us to understand ourselves better in the process, before it’s too late and we lose our humanity?
But to reach a harmonious balance is an idealist’s ballpark. I am not an idealist, and I still think we’re more likely to blow ourselves off this planet with nuclear bombs before we have rogue AIs running loose with incomplete reasons for existence like, “end human violence,” resulting in our extinction. In the meantime, we are in a transitional period that is full of uncertainty, and I’m kind of stressed thinking about who is funding what, and how much any remaining shreds of my privacy and autonomy are going to be decimated by AI technologies. I’ll keep making art to stay sane.
Art is the ultimate utility in my world, and it’s not going anywhere bad. Wen roadmap?
Happy holidays to all who celebrate, whether you observe the Gregorian calendar, or are simply here for all the rebirth of light that comes with observing the Winter Solstice. Whatever you’re into, I’ll see you all next week on my podcast episode with the talented Charlyne who knows quite a bit about rituals.
Oh, also, I’m dropping some nude photography with Manifold starting Monday. Hope you were good, and saved some ETH for it. Merry whatevermas you filthy animals!