AI Generated Art -- Will Actual Artists Soon Be Obsolete?
AI Art Generators Are Fascinating. But As An Artist Myself, I Worry What The Future May Hold.
Art and creativity have always been my escape. On any given day, you can find me drawing a picture of someone’s future tattoo, building a ferris wheel out of popsicle sticks (that actually spins), sewing together my kiddos halloween costume, painting glow-in-the-dark designs on shelves, or pretty much any other artsy/crafty activity that you could imagine. Creating things (including big messes), is what I do best. Such has defined a major part of who I am, for as long as I can remember.
As a kid, all of the school projects that I was required to do, always ended up resembling something that looked closer to art than they did book reports, or science projects. This, you might not have guessed, frustrated my teachers on many occasions throughout the years. Apparently, grading a project that is in no way similar to that of my peers’, nor what the teacher envisioned while designing said project, is not all that easy of a task. Did I care when they would tell me this sort of thing? Not one bit. Had any of them decided that I was no longer allowed to turn in assignments the way I did, it likely would have put an end to me handing things in all together. Art is what made the assignments enjoyable and worth doing. I often struggled with boredom, when working on assignments. Not only was I reading at the college level before I hit grade 5, my report cards never included a single mark below a B-, except in math. Math and I, aren’t exactly friends.
It is this deep connection with creativity that concerns me about AI generated art. Before I actually played around with the technology, however, I didn’t realize just how tremendously it could effect the art community.
It didn’t take much time at all, for me to understand just how serious of an issue AI art generators are, once I began playing around with a couple.
In today’s world, we are already listening to AI generated music. Social media platforms are already using algorithms to censor users. We have even already built an entire response to a “deadly virus”, off of predictive computer modeling —and look how that went.
In many ways, AI and algorithms have already infected all of our lives, and most of us don’t have a clue. When it comes to the tech industry, this is their major moment. Every day, more and more businesses are setting up systems that use AI, and some people are even building entire companies that run off of it. It truly is a moment in history, that we are currently living through, and it is going to change the entire world ..forever.
But when you think about art, the last thing that pops into your mind, is probably mathematics. I'm not saying that math cannot be an art in itself —because it most certainly is. Instead, what I mean is that you likely don’t see “art” as something that would have a mathematical formula. Sure, there are many things that come together to make a piece of art “good”, and that is called composition. The colors, the lines, the space, the texture, the medium, the subject, the style, and even the story that a piece of art tellsn —or the feelings it invokes, and how they are presented/incorporated, are all major players in what makes good art. The most important element, I thought however, is that of the artist’s imagination ..right? I mean, how could something so intertwined with human feeling and passion, ever be created by something as rigid and boring as math?!
That is how I used to feel, that is.
Now that I have taken a little time to play around with a couple different AI art generating apps and tools, my perspective has not only changed, but it has opened my eyes to what the future could potentially hold for any and all creative work.
Am I suggesting that people will cease to make art and be creative? Absolutely not. Do I see many creative jobs being eliminated by AI, and actual artists becoming somewhat obsolete? Sadly, I do. Unless of course, us humans start placing more value in what’s actually real. But due to our insane and unnecessary over use of plastics and synthetic materials, I think it’s safe to say that our priorities will lie in the “cheap, quick, and convenient” realm for the forseeable future. Which means that this just might be the beginning of a dark ages for art and artists worldwide.
I hope such is not the case. Art is important. Real, human-made, passion fueled art. Art is how we express who we are and all of the feelings we have. Without art, life is nothing but a senseless bore. We need art in our lives, and I don’t want to be sitting here saying “I told you so”, in the future.
Even Albert Einstein understood the importance of art. He is quoted saying:
It would be possible to describe everything scientifically, but it would make no sense; it would be without meaning, as if you described a Beethoven symphony as a variation of wave pressure.
On a day-to-day basis, we may not notice just how significant the role is, that art plays in our lives. But that doesn’t mean that it’s role is one we can do without. It is how children learn, it is what brings people together, it’s how we communicate the things we don’t know how to say. And if we’re about to be forced to deal with AI generated art, like we are forced to deal with digitally created music on the radio, it’s probably safe to assume that the quality of the art in our daily lives is going to wind up on the same level as that of the music on 96.7 Chym FM —repetitive, unoriginal, lacking any real flavor, and lifeless.
Sure, AI art generators may have taken the world by storm, and are MVP of the moment right now. But just as Cher's ‘Believe' was a pivot point in the music realm, eventually it will get old. And with the rate in which technology is currently evolving, it will get old fast.
If you are someone who has yet to indulge in the AI art generating world, or simply have no interest in doing so, you may not be aware of just how good AI is at art. In that case, you have nothing to feel left out about. I have just recently taken a few minutes to play around some more with a couple art generating AI Android apps, and I will include some of it here for you to take a look at. But I should warn you that after seeing some of its wonderous abilities, you may feel inclined to check it out for yourself. That is, afterall, what moved me to play around with it myself.
I don’t know what underlying software is behind the AI art generator apps that I use, as I haven’t really read into it. But the two apps that I have spent most of the time messing around with are ‘Dream’ by WOMBO (Play Store link), and ‘starryai’ by starryai (Play Store link).
I have only used the free options on both apps. With starryai, credits are required to generate images, and you get 5 free credits per day, if you claim them. Some options take more credits than others to generate. It all depends on what settings you use. Dream, I believe, offers unlimited generations. Or, I at least haven’t generated so many that I’ve been cut off —and I’ve generated a decent number of images in a day, that’s for sure. Both apps have options that you can select, that the image you are generating can be made in the style of. However, some of them require a premium subscription, and I haven’t used them. Both apps have helpful prompting terms that you can select, to help you generate pictures. I did have a bit of trouble getting the hang of prompting at first, but that’s only because I never actually read the “helpful tips” that apps typically have when you first try them out. I was thinking about watching a Youtube video or two, about how to prompt AI art generators, but have yet to get around to it. It’s something I may still do one of these days though.
But, that is enough detail for now, lets get to the art already.
This is the image that I generated on Dream, by typing in the following prompt:
“a grand ornate beautiful castle, fantasy, hyperrealistic, high detail, full color, hd, realistic, intricate”
Here is another generated with the same prompting, but with a different style selected..
Yep. That was generated in only a few seconds, with the very minimal prompting I mentioned above.
Here are some others..
Again, the same prompting was used, but with a different style selected. The same applies for this next image.
But castles aren’t the only thing I’ve generated images of.
The above image was generated with starryai, and the prompting I input was “Humongous victorian mansion, fantasy, hyperrealistic, highly detailed, full color”.
Above is another castle made with Dream, and the prompting is a little different than what was used to generate the previous castles. This time my input was “an intricate and ornate castle, fantasy, hyperrealistic, high detail, hd, full color, realistic, fantastical”.
The above image was done using Dream, and prompt used to generate it was “an ornate pirate ship with a grand stern castle, fantasy, hyperrealistic, high detail, full color, hd, realistic, intricate”. Pretty impressive, huh?
Here is another, using the same app, the same prompt, but a different style selected..
I told you AI generated art is good.
The image above was generated using this prompt:
“a spooky victorian mansion, fantasy, hyperrealistic, high detail, hd, realistic, full color”
And I used Dream to do so.
Buildings aren’t the only thing that AI art generators are capable of doing well.
Check out the following:
For the above image, once again I used Dream, and “the great mud flood, fantasy, hyperrealistic, high detail, hd, realistic, full color”, was what I used to prompt.
Here is another using a different selected style, but the same app and prompting as the previous image:
Sometimes there are strange artifacts or items in the image, that don’t really make sense, or you aren’t quite sure what they are exactly. But most times, such things only add to the wonder of the image, instead of taking away from it.
Look closely at how detailed the above image is. It probably would have taken me weeks to finish a drawing so intricate by hand. AI however, took merely a few seconds. Again, Dream was used. “Libra scales, fantasy, hyperrealistic, high detail, full color, hd, realistic, fantastical” was the prompting.
The above image was obviously created with a darker tone to the prompting. This time, I input “the end of the world, fantasy, hyperrealistic, high detail, hd, realistic, full color”, and I used the Dream app. I can’t help but wonder how it came up with this image, and why it chose to add the statues the way it did, or what the strange thing is that’s floating in the sky.
Again using Dream, I generated the above image. My prompting was “an astrological wheel portal, fantasy, hyperrealistic, high detail, hd, full color, realistic, fantastical”.
I’ve hardly scratched the surface of AI art generating capabilities, and I’ve played around with the technology many times over the last couple weeks. I obviously didn't include everything that I’ve generated in this post. I only chose some of the best images from my most recent session. There are so many different images that AI is able to generate images of, and as you can see from the ones I’ve included here, the results can be spectacular.
How AI art generators are able to make such magnificent art is sort of like a data-driven type of plagiarism. You see, in order for it to generate art that matches the prompts you have input, a lot of art done by other people is first required. Basically, the AI machine was fed all kinds of images of art that other people have done, and along with each image, was a caption, or tags, that told the machine what the images were of. This way, the machine got a good idea of what a castle looks like, or a cat, or whatever else. So when you type in the prompt, it goes through the data of all the images it was fed when it was programmed, to find images tagged as being the same as the prompt you gave, and it then generates an image based off of that data. In other words, if castles were instead labeled as cats, then when you typed in “cat” for a prompt, the generated image would be that of a castle.
Since the images that were used to create the program, were art that was done by actual people, and was used without their permission, it is a type of digital plagiarism, however.
Then again, I am one of those artists that have always argued that it is also plagiarism to be inspired to draw your own version of someone elses art. But I have never cared about artistic plagiarism. To me, there is a difference between imitation and inspiration. One is the attempt to recreate the same, the other is creating a different version. With AI art generators, however, I do feel like it’s a different type of plagiarism, due to actually having been programmed with the art people have created, without the consent of the artists. Since the way it works is strictly data-driven, it is therefore generating images using exact parts of art that was created by humans, and money is being made using people’s art, without their consent. So, I feel like it’s wrong.
Obviously I don’t feel too terrible about it to not play around with it though. So I really don't know what to think, in the end.
All I know, is that real art made by real humans, with real feelings, will always be more valuable to me in the end. Regardless of how incredible good AI art generators are at putting together data to make an image.
That’s all I have to say for now.
Thank you, and please don’t hesitate to comment below if you have anything you’d like to say, or have any questions you’d like to ask.
-Davis