masterpiece. High quality photorealistic image. Medieval setting. A beautiful naked slave woman is auctioned off to a (((huge group of leering men and women))). Her hands are tied behind her back and she is ball gagged. A (((fully clothed aucitoneer))) stands beside her, pointing at her. The slave woman and the auctioneer are (((standing))) in a wooden wagon parked on an unpaved dirt road. Bondage, rope. Handtie, Hands bound together with rope, hands tied behind, wrists crossed, wrists parallel, palms out, boxtie, showing hands, prostration, HDA, gag, ball gag, ballgag, ballgagged.
I got an AWFUL lot of images like the one above from that prompt using the Cyberrealistic Cyberillustrious Pony model. What the hell was the AI art program thinking? Why does it HATE medieval settings so much? I mean it's right there in the front, the first description of the image's contents, and what do I get? Pimp Daddy in a three piece suits and four women who would not look at ALL out of place in a modern Amish farm or bake sale or whatever they get up to in Amish communities.
What the prompt DID get was the wooden wagon and the unpaved dirt road. It got those COLD. They weren't exactly ESSENTIAL to the image. Especially since the AI Art program NEVER put the slavegirl in the wagon.
The triple parens are supposed to give the AI art program the idea that the items inside the parens are important and should be given priority. (And yes, in another iteration I did put "medieval setting" in triple parens. No dice.) I did have some success. Before I put "standing" in triple parens Civitai kept giving me pictures of naked men and women kneeling in the dirt, or fucking in the dirt. Like this:
That's why I love AI art. It's like a box of chocolates... you never know what you're gonna get.


