Select Page
AI image generator by Jean Giraud --v 6.0-U3 - Variations (Strong)-U2

Click the above image to enlarge.
There’s also a link to a Flickr album containing all the AI images I uploaded.

Should a modern artist use “AI image generators”?

Is this a tough question for you?

“It’s unethical for an artist to use one because everything they produce must be original. They are nothing short of plagiarism.” That was my mindset, so I couldn’t allow myself even to try one of the free image generators!

Also, some produce strange errors in the images. Like humans with missing fingers, or hands. So I thought “you can always tell a generated image.”

Note: By “AI image generator,” I mean a text-to-image app that was trained on images. They produce completely original images.

“A good artist experiments and explores…” — Shane McGeehan

Then one day it hit me: if I have original ideas, this might be a quick way to “brainstorm with myself”. Possibly create some concept sketches. I’ve studied anatomy enough that I can fix any errors and improve on what the machine did. I could also “mix and match” by using the best results from each “image prompt”.

When you think about them like that, it’s no different from using reference photos you find online! Or, if it is different, it’s better! Because it’s an “original reference photo“!

Free AI Image Generators

When I finally looked up “free AI image generator“, I found several good ones. A good free one is called Koala, which uses Stable Diffusion. (Read about it and see results here: KOALA. There’s a version you can run online, but it’s quirky (if you’ve not used Iron Python or Jupiter Notebooks for Python). To use it, just hover over each of the “[ ]” in succession. You’ll see a “play” icon. Click it to run that command, then go to the next one and continue. Edit the “prompt” line and click run to get your own images. There are only a few clicks to use it.)

Koala produced some encouraging results! For concept-brainstorming, it would be great. I thought I might even use some as “sketches” for custom 3D models, which I use in my computer and VR game. (A series named FlyTime, which is under development. There are some early FlyTime screenshots on Flickr. There’s a lot to do, but I’m determined, so stay tuned!)

There are a lot of other choices. The free ones don’t look that good to me. There are several others. (Including OpenAI’s DALL-E. They charge by the image. So I’ve not tried it.) Yes, DALL-E and some others produce pleasant images. So read on!

The best subscription AI Image generator I’ve seen.

Naturally, I got curious about other generators I’d found in my search. One of the top results was a website named “Prompter Guide“. Which was created by Shane McGeehan, a friend of mine on Facebook! (Is that strange??) His results are amazing! He’s using Midjourney. Check out his eye candy at PrompterGuide.com!

How much does Midjourney cost?

Midjourney requires a subscription. (No free trial.) Their starting price is $10/month! (Subscribe yearly and pay $8/month.) For that you can generate literally thousands of images per month!

So first, I thought up some ideas. Then I finally talked myself into it.

Midjourney is so amazing that I’ll probably go with the yearly subscription to save a couple of dollars each month. You can find Midjourney documentation here.

Philosophical questions for AI

What if you use a prompt that contains only adjectives, personal pronouns or philosophical ideas? What would you get? Would it look related? (How could that be??) Then it hit me:

What about prompts including “AI”. Surely it could give me some sci-fi images! Right?? Soon after I thought about that, I thought, What if I give it a prompt for “AI image generator”?!? Would it “know about itself”? Or would it combine some sci-fi stuff to create something new?!? It turns out the answer to these questions is “yes“! 🖖😎

Here’s an “AI Image Generator in the year 2300 by Jean Giraud.” You may know Jean Giraud as Moebius.

Click to enlarge, or see the this album:
AI Image Generator creates “AI image generator”! How does it see itself?!?

I created these on my first day of subscription! I used PrompterGuide, but I provided artist names I liked (and he doesn’t have listed). I also used some of his preset selection for “futuristic glow”, etc.

The image at the top and bottom page is one of many favorite results about “AI on AI”, or “Existential AI”! Midjourney produced this when I used “Jean Giraud” as the artist and “AI image generator” in the prompt. Click the image to enlarge size. (Or see it and more in the Flickr album links on this page.) This is a “creative variation” of one result:

AI image generator by Jean Giraud --v 6.0-U3 - Variations (Strong)-U2

AI Image Generator in the style of Richard Corben.

Update 2024-03-23:

I burned through my 200 images in the basic Midjourney subscription (with allowance of 3.4 GPU, or ~200 images) hours in < 2 weeks!!

Should I be pickier about my prompts?? Should I upgrade my account any pay more each month? Is it more economical if I just “top off” my subscription hours, at the rate of $4/2 GPU hours? Or maybe both?

Stay tuned for the answers to these and other intriguing questions!
Same Bat-Time, same Bat-Channel!

Midjourney AI image. The full prompt is found in the image name, and in the description. Just click any image (on this page) to view it in gallery mode and see that.

AI Image Generator for “Bat Channel” and Batman logo on clouds.
Click to enlarge and see the full Midjourney prompt.

All AI-generated images I’ve uploaded are found in

“my AI Generated Images” album on Flickr.

Share This