Sci-Fi World Building

AI into the sketchbook...
[Using image-based AI to improve existing artwork and concept pieces]

One of my problems with the current state of AI generation is control. In an effort to take control of the tools, I started using image based AI (with prompt support of course).
"Anyone can write a prompt." It's true. But as an artist, having control of the tools is optimal to getting the desired results. I've been experimenting with taking finite control of the process, and having greater accuracy in the outcome by literally giving the thing my own art

Here I have repurposed old drawings and thumbnails from my sketchbook using Stable Diffusion to add new dimension to pencil sketches and artwork. I could have accomplished this all in 2d/3d, but it would have taken me weeks. Using previz concepts, I can "kit-bash" the AI by providing the correct pieces in advance. This is my first experiment into incorporating AI for efficiency in production. This workflow is valid for everything from concept prototyping to rapid processing of storyboards. My next experiment will be translating mocap and animation via AI using the same base concept: Steering the outcome by providing solid reference- that I've designed

Image based
SDXL 1.0 and StableDiffusion 2.1
(Prompt and thumbnail sample in comments)