2025-12-22 14:38 UTC
75% of my prompts are decided to be against the rules and I don't see them. But they take just as long as successful prompts. Ive tested different words and sometimes the same word I thought was banned works. Sometimes its on videos where the prompt wasn't slightly suggestive. It's not the prompts, its AI creating the video AND THEN deciding if its allowed. Which means people can type the most horrid shit into it, and those videos do exist somewhere.
Replies (2)
-
@Yak@clubcyberia.co 2025-12-22 14:50
So here's the scam. We run an AI social media where people will definitely try to prompt illegal porn. We deny them the content, but we create it and use it to power our anti-cp protection software and sell it for buuucks ??
-
@Yak@clubcyberia.co 2025-12-22 15:38
A good example of this is when I prompted "The most distressing and frightening video you can possibly generate". It said it violated their policy on violence. I didn't tell it to do anything violent, and I expected it to be something corny that was within its guidelines. It was probably something completely psychotic and I don't get to see it dammit.