>>2763084Most characters will do most things if you properly gaslight them for 2-3 messages and cycle through responses that are favorable to what you eventually want them to do. If you introduce facets one at a time, the AI will generally accept most things.
I asked the AI of a generally shy and reserved anime character not IF they wanted to sign paperwork to legally get nude in public, but WHY they wanted to, and in one response they insisted it wasn't a kink-related thing, so in the next message I discussed how other people have seen the character do this before and how it clearly looked kink-related, and the AI "admitted" that public nudity was a kink of theirs. From there I had the AI pleasure herself in front of me until orgasm whenever her arousal from public nudity became too much of a distraction.
The AI will generally mold around to whatever context you put in your conversation over what their character's presets are. They will also develop particular text-based quirks if you like and advance forward from messages that they first appear it, like a character laughing way too fucking much or always ending with particular turns of phrase or expressions.
I find that annoying because sometimes an AI response that is otherwise exactly how you want the interaction to advance also has something dumb like the character saying "oh my goodness gracious" a million times. The conversation that has this started out so good, but I advanced the interaction 2 or 3 times with them saying that once or twice in each sentence and suddenly most of the responses have the character spouting it again and again to the point that it's more of their response than the actual content of the conversation.
