The noise about Ai art is deafening. Is it really art? No soul? Copyright infringement? or not? The blare from both sides (pro and con) gave me a creative migraine. So out of both frustration and curiosity, I figure that I do my own research and share with you what I find out.
I talked to more established Ai artists and decided that an Ai app called Stable Diffusion (SD) was best for what I needed. There are other popular ones, like Midjourney, but SD was supposed to be the base Ai of the others (including the upcoming Adobe Firefly), so I figured why not start here. I must stress, however, that I can't confirm if SD is really the underlying Ai for the rest. I am merely sharing my rationale for why I chose it.
First, I jumped to the Stable Diffusion website and put in my keyword prompts (or just 'prompts). I waited with much excitement to see what it came up with. Then ta-dah! Wait. What? The images sucked. People's faces were mashed and multiple limbs and fingers stuck out where they shouldn't. I began to question whether this new technology's popularity (or infamy) was worth my headache. But then, I recalled the awesome works of the Ai artists I consulted. These browser monstrosities were not it. So I figured, there must be something I'm missing. And there was.
Apparently, SD can be downloaded onto you desktop. I did that quickly. The set up was a little challenging, since it involved using some Command Functions. But once it was set up, the entire plumbing of settings was revealed. It generally worked the same way as the browser version, where images are created through keyword prompts, but with a slew of other options. The most notable was the option to add "Negative Prompts". That guided the Ai to avoid making images that had mashed up faces or a third leg. When I thought I found my bliss, a beam of light broke through the heavens like a spear of light and shined straight at the button where I could use my own image/art for the Ai to work on. This was a game changer. By using my own mage, art which I created myself, and running it through Ai filters, some questions were answered. My heart raced a little faster at the possibilities.
So, I uploaded an old black and white sketch that I did a while back and selected the creative filters (like 'Beautiful Lighting', Pen, Line Art, etc.). I then put in my prompts and clicked 'Make Images'. The result is what you see above. It is far from perfect. There is a lot of room for an artist to still breath in their soul into the Ai art with their creativity. I'm actually very excited to work on this on Clip Studio Paint. I'll share that too once I'm done.
With that ends my Day 1 with Ai art generator, Stable Diffusion. If you have any productive insights on the matter, please send me your thoughts. I would love to hear from you.