AI art

AI art generation with single word prompts

Can decent AI generated art be made using stable diffusion and single word prompts?

If you’ve experimented with AI art generation, whether using stable diffusion or some other system like Midjourney, you’ve probably noticed that a carefully designed prompt can make or break the outcome.

The golden towers of anor londo at sunset with god rays breaking through thick clouds, 8k resolution concept art by Greg Rutkowski, Artgerm, WLOP, Alphonse Mucha dynamic lighting hyperdetailed intricately detailed Splash art trending on Artstation triadic colors Unreal Engine 5 volumetric lighting

This can lead to some complex prompts that not only describe the subject of the artwork to create but also the styles and parameters for it to attempt to follow.

photorealistic portrait of a steampunk cat wearing a wizard robe and pointed hat by Darek Zabrocki, Gilles Beloeil, Cedric Peyravernay, Jordan Grimmer, Tyler Edlin, Greg Rutkowski, , hyperrealism, photorealism, trending on Artstation

But… what happens if you give the AI a prompt containing a single word? No style guides or anything else, just a single word.

I’ve got a stable diffusion install on a machine here and it’s a bit slow but I’m going to give it some single word prompts and see what it comes up with.

I’ll try to give it some that it can’t mess up, as well as some more abstract ones, or ones that could be interpreted in multiple ways. Results below… but be warned… you might not like what you see. 🤣