I've started playing with AI images (Tensor.art). Here's 3 that kind of work.
The AI seems to be very good with ethnicity, age, skin tones, expressions and body shapes. Not so good at posing, camera angles, nails, whips, ropes, chains, and everything else. And it won't put the girl in the fire

I used Juggernaut XL with prompts like :-
"25 year old gorgeous Puerto Rican girl crucified, tanned,large_breasts,suffering,crying,athletic,sweating_profusely,sweat_drops, at_night, fire, flames"
and negative prompts of :-
"Jewelry, makeup, paint, clothing, belt, bikini, bra, panties, straps, Malformation in fingers, deformed fingers, ugly fingers, Six fingers, duplicate fingers, duplicate navel, double belly button, ugly belly button"
Not sure how much longer it'll be before the novelty wears off, but it's fun for now
