Hi Blogleech! I’m a CGI university student, and I think I can offer a unique perspective to this debate. A perspective that’s 1480 words long…. Jesus Christ I didn’t realise just how bad I needed to talk about this. 0.0
First, I want to say that I 100% understand where professional artists are coming from. The way AI art is being marketed to both artists and the layman, makes it seem like there is no artistic vision to the process.
For anyone who isn’t an artist, let me tell you that artists are extremely pedantic. A piece of art starts as an idea, not just of a prompt, but as a vision (an artistic vision you could say). The pose, the environment, the colours, all of that is in our head when we pick up a pencil.
So, when AI art dropped and took the artistic vision OUT of the artist’s hands, it was terrifying! No longer was the artist able to control a piece, instead, you type a prompt and let the AI do all the fun stuff for you and that’s just, so, disheartening.
So last week, I was 100% against AI art. It added to my burn-out because it felt like nothing I was drawing mattered! I got major anxiety and I couldn’t sleep at night. However, one of my coping mechanisms is to look into what I’m afraid of, since it’s often the unknown that scares me the most.
So, once I headed home for winter break, I took a dip into the AI art world to see what all the fuss was about and guys… artists and, by extension the artistic vision, aren't going anywhere.
You see, most of the artworks you’ve been seeing form AI art have been using a discord extension called MidJourney. This is the software that AI generated comic was made with. You add it to your discord server and type prompts to generate art. It creates 4 small, blurry images first so you can choose which one you want to expand upon. You can retype your prompt with slight changes to improve the smaller blurry image until you have something you like, and then tell the program to export a bigger final version.
Sounds impressive doesn’t it? Well, what if I told you that this is actually shit. First of all, it’s subscription based after your first 25 prompts, and I think we all have similar feelings when it comes to subscriptions. Secondly, it can ONLY produce finished images!
“But ShinyDino, aren't finished images the only thing AI art can make” NO! Because there are more software in town and, with the right talent, can outcompete MidJourney at its own game.
Behold, STABLE DIFFUSION! Now if you were to play around with the program for a minute, you’ll quickly realise that this software is just a shity MidJourney. However, this shit has an advantage that MidJourney lacks: It’s open source. Open source means that the internal code is open for anyone to copy and edit, meaning that coders from all over the world can build upon it and innovate like crazy. Which lends me it the program that I’m MOST excited for: AUTOMATIC1111
AUTOMATIC1111 (Yes, it’s name is in all caps) Is something that I like to call “Photoshop 2.0 ^ 10000”. It’s an interface for Stable diffusion that operates in your browser and it lets you do 2 things that I think will bring the artist back into AI art: Image to Image, and Inpainting.
Image to Image, unlike text to image, converts an image, sometimes as basic as MSPaint, into a more hi-rez image with a chosen style. It still requires a text prompt, but that’s more so the computer knows what the fuck it’s looking at and it can sometimes interpret what to do anyway.
Meanwhile, Inpainting is a function that lets you mask out a part of an image and re-roll your prompt, letting you fix mistakes in the image generation, like extra legs and wonky eyes.
Now dear reader, I want you to partake in a thought experiment: It’s the distant future of 2024. You’re in charge of employing new concept artists to the studio and you have two digital portfolios on your laptop. Let’s say they’re on Artstation, just to keep it topical. The first portfolio is from an artist that uses MidJourney. Their page is PUMPED with art, the definition of spam, all very pretty and with different styles, all made in under a couple of minutes.
However, on the second portfolio, there’s a show-reel. It’s only a few minutes long so you open it up. You watch as the artist makes a quick coloured sketch. They then export that sketch into AUTOMATIC1111 and see the sketch get converted into a fully rendered illustration. But that’s not all! The artist then fixes the AI’s mistakes using inpainting, then exports it back into photoshop, adding post processing and effects and other little details that make the art pop. The video ends, and you scroll down, finding the final illustration, more detailed than any other concept art you’ve come across.
Now tell me reader, who’s the employer going to hire? The guy who can type nice looking prompts, or the artist who knows both art and their software back to front, who can not just use the styles already provided by AI, but can adapt them and make them their own, and can edit their own images in real time to match both their own, and their directors vision.
IT’S PHOTOSHOP GUYS!!!!!!!! SUPER SONIC SPEEDY PHOTOSHOP!!!!!
Artist + AI tools = an artistic monster that text to image propters can’t compete with!!! No longer do I see MidJourney as a threat to artists: I see it as the newest version of clip art! Something cute that can satisfy the layman, but not useful for the industry. It’s the artists, those that can build upon AI art that the industry needs.
“But shinyDino, does that mean that people will lose jobs if they can make more art in less time?” When, good reader, has that ever happened? NEW THOUGHT EXPERIMENT!!!
You are a producer in an animation studio. Watching over the departments, you notice something about your background painters. Their productivity has exploded over the last 2 months, making an average of 12 backgrounds a day instead of 3, despite the fact that they’re taking longer lunch breaks. You get them into your office and ask them to explain, and they teach you about AI art and how they’re using it in their workflow to speed up the process.
Now, dear reader. You have two options: The first is to fire all your background staff except for 2, and finish the studio’s project while saving money. This IS an option, especially in a financial recession. HOWEVER, if you do this, then you’ll be shooting your studio in the foot.
Here you have employees who aren't being replaced by technology, but are using and abusing it for their own gain. You don’t fire these employees, you use them to compete against other studios! Make more impressive backgrounds! Wow the competition!
As Sony was innovating in technology, did they ever stop and say “Welp, good enough! Let's cut our art departments so we can put money elsewhere”? NO!!! They continued until they made graphics so realistic, basement dwellers finally learned what women look like!
Which brings me to the thing that I’m most excited for: I truly believe that AI art is going to bring us a “Visual Storytelling Boom” similar to the rise of the Indie game in the 2010s. Because AI art gives artists something they’ve never had before: Time.
Hey artist! Have you ever had an awesome project pop up in your head that’s just too ambitious? Of course you have! 20 popped into your head the moment I asked that question. Well, that project is no longer ambitious. You can render an illustration in 10 minutes. You can make a comic page in a day.
An artist called realspacejunk on Tumblr recently released this illustration. They say it took half a year to create because of the insane amount of detail in the image… Now imagine that taking a week. Imagine a TV animated series in this style.
All those kids that grew up on Avatar can finally make the long form series of their dreams! CGI artists can render backgrounds in a similar vein to Arcane and focus on what they actually love: the Character Animation. As professionals make new gorgeous and complicated styles, the old and simple ones will be available to the everyday hobbyist, like how 8-bit is the go to style for indie-games.
Next year, we are going to see a visual storytelling revolution! And you know what I’m scared of? What’s now keeping me up at night? The fact that art professionals are being scared away from using this new tool. It’s heartbreaking! I don’t want tech bros to be heading this revolution, I want the charge to be led by artists!
So please, Blogleech, if you find any of this convincing, spread the word. Spread the links and tutorials, the technology, cus I’m a student who’s wacom’s just died and waiting for a new one for christmas, but you have a platform! This is not the end of artistic vision. It’s just the beginning.
