Lately, I’ve been flirting with art, in particular with AI generated art and character designs. Because, of course, I want to draw my own comics someday. It’s a ‘greener’ pastures situation, because when words are hard to put down… why not just draw them?

Well, with AI Generated art, you use words to create art.

I am fascinated by it. Yes, there are a lot of concerns about AI Generated art, but I see a lot of benefits. If I can create actual pictures with my words, that would be amazing. Just take the images from my head already!

(What I don’t like is the idea of the AI programs stealing. That the programmers did not think about how they trained their AI and where they got the images to train it. 

If the AI I’m using was trained by art stolen from artists, that sucks. I don’t want to be a part of that. The original artists should be compensated for their dedication to the craft and their hard work in creating a unique style. Instead, if it was stolen from them and can now be endlessly recreated using a computer. I hate the thought I’m contributing to such a mindset.

Artists learn by imitating other styles. They evolve as artists by learning from others, finding teachers to guide them, and finding art that inspires them. They develop new techniques and master the old ones. You train your eye to see shadows and light. Using computers to create is a part of art’s evolution, but artists must be included in the process. Their work acknowledged. Compensated.)

I wanted to use AI to make a book cover.

My first attempt at using AI generation was with Craiyon. It models art directly from written prompts. I gave it this prompt: a silhouette illustration of a path winding a forest of black thorns.

This was my result. Not bad. It’s fine, but not really eye catching, not for a YA adventure romance book.

I’ve also used an AI generated art program called ArtBreeder.com. This website makes gorgeous portraits, like the kind you’d find on a profile. Here are some examples I created of my Fairy Doctor characters. There is Marquez, Lillia and the Baron. (Not the definitive versions, those exist only within the novel, and in my head and my readers’ heads. This is just a good idea of how I imagine them visually.)

I also created Lucianna from The Thorn Path too. I found an image to start from, then evolved her using another, then tweaked the settings for coloring, nose, more Black ancestry, a little Asian and Latino, younger, etc. I like having these options. It makes me feel like I’m inputting and creating art.

The beautiful thing is you can look back and see the DNA of where an image came from, all the portraits used before to create a new one.

Once satisfied with my portrait of Lucianna. I downloaded the image, and I put it into Art Breeder’s Outpainter.

I added the keywords: profile of a girl in a dramatic forest.

BAM I had a cover art.

It wasn’t perfect, though.

What was up with the shadows on the left shoulder and such odd-looking clothing? Is she wearing anything under that coat? AI art is often this way. It just resembles a picture, but isn’t created by someone who can see. A hand can have six fingers, because as long as it resembles a hand with fingers, that’s good enough to the AI. Uncanny valley does not exist for AI. 

So, to fix this I downloaded this version, and I put it in Illustrator. This way I made my book cover the right size, 1500px x 1500px. And I could export it at 72 dpi. I added this spooky tree effect.

And made the background fade to black, adding a slight vignette. Then I painted over it in Procreate.

I made her face less pimply. I took away the strange shine on her forehead. I fixed the collar of her coat, so it would look like she was wearing a watercolor forest-green-camouflage coat. Below is the image I uploaded to Kindle Vella. It will become my first book cover.

The trick will be the second book. How can I use the same AI generated character to make a different book cover? Oooh, tricky. I’ll keep you posted. For now, I’m only on episode 4 of the Vella story. 

Have you checked it out? I’ll be posting more often to promote the story and will let you know how it all turns out. (I’m also slowly working on Fairy Doctor Book 3, it will be coming out this year. Promise! But revisions are slow and I’m impatient writer/artist. New story, yay!)

I had fun making the book cover for The Thorn Path. It still involved a lot of design work, from the initial creation of a character to repainting what the AI created, fixing shadows and colors. It was collaboration between me and the computer. It took half a day to make and I’m satisfied with the outcome.

What do you think of AI generated art? Anything you loved about it? Do you hate it? Are you an artist using it too?