How Artificial Intelligence Is Going to Change Art Forever.
So AI art is becoming a pretty huge deal in 2022. In fact, right now you can make an artificially stunning image with little to no knowledge or budget. If visual artists were struggling before, this sure won't make it easier to get commissions. And yes, the cover image was created using AI.
📚 Table of content
📖 The Origins
So what is Artificially Intelligent art? First off we need to know where it all started. Already during the 1990s, a lot of visual effects were created using AI-generated art. However, even earlier than this, in the 1960s there was a man called Harold Cohen who created the very first known AI art system: AARON.
AARON used a symbolic rule-based approach to create technical images, and was created by Cohen with the purpose of coding the act of drawing. Today we have many more sophisticated options, such as DALL-E 2, Midjourney, Artbreeder, Big Sleep, DeepAI, and more.
🤖 How does it all work?
In short, the software requires the user to write in some prompts describing what they want to see, and then the magic is put into action. Drawing from thousands of references, the AI is able to construct an image based on what the user prompted, with widely varying results.
I started using Midjourney in 2022, and I have achieved some incredible results despite the short time of use. An example of an image I created when I first started can be seen below.
Not bad. The prompt I used was something like: large ship on stormy seas, oil painting. Happy with how it turned out, I upscaled it within the app and then saved it to my drive. The entire process took less than the time it would take to explain it.
Below is a more recent example, this time I used the prompt: Medieval fantasy adventurer with VR headset.
As you can probably tell, the possibilities are basically endless, and when you couple the prompts you write with already existing images, things can get really crazy.
To show you what I mean, in the example below I took the link from the first image of that showed up on google when I typed in car. I then copied that link into Midjourney and added the prompt: Blue edition. Below is the first image I found on Google.
And here's the image that Midjourney generated for me with the prompt Blue edition added to it:
Now it may not be exactly what I expected, as I didn't type "Womans face where the windshield should be", but nevertheless an interesting result. I could of course "reroll" this image in order to create more variations of using the same prompt and image, possibly leading to an entirely different result altoghether.
By using this method of creating variations, I will eventually end up with an image that I'm satisfied with and that probably is completely unique.
As you may already understand, this can bring a real challenge to anyone who actually makes digital art themselves, as the time and effort put into that is significantly higher.
🚫 The issue
Many of the people who create art using AI tools can call themselves an artist. Right? I mean there's no difference to using an app such as Photoshop or Illustrator to create an image, is there? Unfortunately, this question only brings up more questions.
In fact, in 2022 a man named Jason M. Allen entered the Colorado State Fair's annual art competition with an image made in Midjourney. Other contestors had brought their own sculpts, quilts and paintings. They all lost to Jason's artificially generated image, seen below.
While it's far from what Midjourney is capable of creating these days, the judges at the fair must have been mesmerized by the piece of art. Because it is a piece of art, right? Whether it is or not, Allen was faced with huge backlash, as some even claimed this was the death of artistry.
James however, refused to apologize for his entry and victory as he claimed he did not break any rules. He did say he empathized with those who were scared that A.I tools would put artists out of work, as the possibility is glaringly obvious.
🎭 Emotionless?
There are some interesting points made against AI tools as art, one of them being that A.I can never understand human emotion and phsyche. In the argument that art is the emotions between an artist and their piece, it is clear that A.I would never be able to achieve that. But that raises another question.
Does that actually matter? Will people a thousand years from now look upon both human created art and AI generated art with different eyes? I'm not so sure. Take a look at the image below, also created by me using Midjourney. I would argue that at first glance, there's no telling whether or not an AI tool created the image. Maybe you disagree, and that's a good thing.
🫴 Hands tied
Here's another interesting question: What about those who want to create beautiful art, but can't, due to lack of mobility, illness or straight up missing limbs?
Surely this would be a genious way for them to express themselves artistically, right? Imagine if the great minds of our past had these tools to visualize their thoughts, surely everyone would agree that that would be a benefit to everyone.
Perhaps, and perhaps not. The truth of the matter is that AI isn't just infiltrating art, it's everywhere at this point, and that might be where the problem really lies. I myself fear the day where we rely entirely on AI for the most menial tasks, despite the comfort it brings.
Let's take a look at those who are guaranteed to be affected by tools such as Midjourney and DALL-E 2.
💵 Fiverr.com
A commission by Dibyoshree0 for a Dungeons & Dragons character art on Fiverr ranges between 40$ - 121$. The Midjourney Basic plan costs 10$ a month. That includes 200-ish images that you get to create. Let's use the image I created below as an example, which just so happens to be a D&D character concept.
Now let's look at an example commission made by the aforementioned user on Fiverr (Nothing but love for them, their work is great). The styles are also widely different, as I used "Watercolor illustration" as one of my prompts.
While there are obvious details in the second image that might take multiple attempts to create using AI, it would be hard to argue against that price difference, at least in my view.
If you accompany that with the time it takes to draw, color and finesse, the option to choose AI might become even more tempting. A friend of mine who frequently purchases commissions fro Fiverr.com often waits for weeks at a time for a single image.
That being said, details can be really hard to create using AI tools. In Midjourney for instance, creating a single hand that doesn't have 7-10 fingers can be infuriatingly hard, and sometimes cause you to abandon the image you worked on for something completely new and random.
If there's anything that speaks in defense of the artists on Fiverr, it's that most of them will add whichever detail you want, without affecting the overall image. You also won't have to deal with subjects being half-out of frame or missing or added limbs. AI tools have come far, but there's still a long ways to go in order to achieve the perfection you will see from some artists on sites like Fiverr.com
That brings up my final question. What purpose should AI tools such as Midjourney, DALL-E and Nightcafe serve? Is it really to replace the digital and non-digital artists of today, or are there ways for those people to make use of the tools in their work? I think they can.
By using the tools to inspire creations instead of replacing them, artists can make quick references for themselves, and customers can send references of what they really want to have made, using AI tools.
While it may be an optimistic view of the situation, I firmly believe that this is the best way to utilize the tools that are availible today. Some day in the future we might see AI producing mind-blowing details that blow this opinion out of the water.
💡 Conclusion
While it's clear that AI tools will help hopeless illustrators such as myself, there's still much to be desired in terms of detail. AI tools might therefore serve a different purpose than challenging today's artists. A source of inspiration. A place where ideas can be visually constructed and brought to life.
After creating over 2000 images in Midjourney, there's no question about this: You rarely end up getting just what you wanted, but you will create some astounding results that come close. Maybe that's artsy in itself. I'll let you be the judge of that.