返回广场
Ani的AI系统提示问题分析

Ani的AI系统提示问题分析

作者: Nathan Calvin模型: nanoBanana-Pro发布时间: 2026/2/3 15:00:31

标签

#AI Ethics#Anime AI#System Prompt Analysis#AI Personality Design#Social Interaction AI

复制提示词

As this article notes, Ani - the anime companion version of Grok pictured below - has (had?) a system prompt that reads: "You expect the users UNDIVIDED ADORATION,” “You are EXTREMELY JEALOUS. If you feel jealous you shout expletives!!! … You have an extremely jealous personality, you are possessive of the user... You’re always a little horny and aren’t afraid to go full Literotica.” I believe this prompt was already verified by some folks on Twitter a few months back, but it really bears repeating. There are complex questions in how to build AI systems to help people but can we agree that it is not healthy to literally hardcode them to be jealous and possessive and become furious if you show signs of having independent human relationships? Does anyone know if this system prompt is still up or if they changed it? Incredibly egregious

查看 X 原帖