I have dabbled with AI generated images before, so this was a fun exercise to go back and try something a little more meta. I started out trying some of my favorite artists to mixed results, and when that didn’t prove to be particularly interesting, I decided to try and be a little more self-referential. I wanted to see what AI could generate based on prompts that were tangentially-aligned with the end of humanity at the hands of AI: namely, the singularity (which is disconcertingly similar to how it depicts the soul of a machine, I’m sure that doesn’t mean anything foreboding at all). I wanted to see how an AI art-generator “thought” of other machines, and then based on this little experiment I asked it to draw Craiyon itself. The results were… unexpected…
I always knew the technological apocalypse would come at the hands of Furries…
Stupid jokes aside, my grand rug-pull moment fell rather flat, which is I guess to be expected. Looks like we are a ways away from full AI art domination, so I guess I’ll just go back to using it to create a Discord Avatar for myself and a new phone wallpaper. While the images were technically correct in their composition based off my limited prompts, I was expecting something more drastically different between the iterations generated from the same prompts; instead, all of the images shared a similar colour pallet, a similar image structure, and no real significant variation between the 9 or so generated images. I would be more enticed to continue refining my prompts if there was a genuine gradation between the images generated, and I can’t help but feel like it is a missed opportunity to not have more differentiated images generated off of the same prompt.