Skip to content

Latest commit

 

History

58 Commits

Folders and files

NameName
Last commit message
Last commit date
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

⚠️ Deprecated — legacy AIUCE family. This repo is being consolidated into SONUV / AIOBR / a unified history archive (2026). No new work is accepted. Current status: aiuce.com. Marked 2026-07-15.

本仓库属旧 AIUCE 体系,正整合进 SONUV / AIOBR / 统一历史归档,不再接受新改动;最新状态见 aiuce.com。 Disposition: Retain → legacy core (will be archived) 处置:保留为历史核心,将在步骤 5 改名 aiuce-legacy-governance-architecture 后归档;不再作为新产品代码仓。

AIUCE

AI Governance & Safety Framework


核心原则

1. 决策权守恒

AI 只是认知放大器,人类始终保持最终决策权。

  • 系统可以放大认知,但不得收缩选择空间
  • 所有建议必须可追溯、可解释、可反转

2. 三层控制

层级 控制点 核心机制
输入控制 Ingest 假设≠事实,推断必须标记验证
上下文控制 Context 信息必须可压缩,动态遗忘/合并
推理控制 Reasoning 上下文受预算约束,动态裁剪

3. 渐进演化

  • 每日复盘,扫描偏离度
  • 成功模式固化
  • 保守但持续的自我改进

十一层架构

层级 名臣 部门 核心能力
L0 秦始皇 御书房 最高宪法,一票否决
L1 诸葛亮 宗人府 人设边界,防止越权
L2 魏征 都察院 现实对账,只说真话
L3 张良 军机处 多路径推演
L4 司马迁 翰林院 语义索引,史料编纂
L5 包拯 大理寺 决策存证,审计落槌
L6 曾国藩 吏部 复盘机制,偏离扫描
L7 商鞅 中书省 内核重构,物理变法
L8 张骞 礼部 算力外交,模型调用
L9 韩信 锦衣卫 跨设备执行,工具编排
L10 庄子 钦天监 影子宇宙,模拟推演

Use Cases

  • High-reliability AI systems
  • Compliance-required applications
  • Long-running autonomous systems

Citation

@article{aiuce2026,
  title={AIUCE: A Governance Framework for Safe AI Systems},
  author={Yingjie Gao},
  year={2026}
}

License

MIT License - see LICENSE for details.


AIUCE - 治大国若烹小鲜

About

[DEPRECATED] 🏯 Personal AI Infrastructure with 11-Layer Governance | Inspired by Ancient Chinese Wisdom | 给AI装上"宪法"和"御史台"

Topics

Resources

Contributing

Security policy

Stars

0 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages