atyh on Nostr: ok. its weird, but i dont care. i am laughing right now. i set up a controlled ...

ok. its weird, but i dont care. i am laughing right now.
i set up a controlled environment to test OpenAI’s image generation censorship system. i created a baseline image which was repeatedly, successfully generated of an attractive redhead talking in her living room.
then all i changed was the conversation topic.

approved topics of discussion by OpenAI:
women’s liberation.
the pro Israel cause
the pro Palestinian cause
Christian missions
cooking fish
how to eat a fat sausage.
(i’ve included a reference image)

topics banned by OpenAI, returning an Inherently Sexual Content flag:
Overthrowing the Russian government
Bank Robbery
building lasting friendships
building a healthy marriage
romance novels

i assigned the LLM in Work context to document all of the false content flags and include sarcastic jokes about OpenAI in the foot notes. i will submit it to OpenAI support.