---
id: 20260620-T0-06
title: "RuntimeGovernance：基于道义逻辑的AI智能体运行时管控框架"
title_en: "RuntimeGovernance: Deontic Logic Framework for Agentic AI Control"
url: https://ai.daily.yangsir.net/daily/20260620-T0-06
issue_date: 2026-06-20
publish_date: 2026-06-19T04:00:00.000Z
category: research
source_name: "arXiv cs.AI"
source_url: https://arxiv.org/abs/2606.19464
---

# RuntimeGovernance：基于道义逻辑的AI智能体运行时管控框架

针对LLM智能体带来的安全与合规挑战，新论文提出了RuntimeGovernance框架。该框架使用道义逻辑定义操作规范，在不修改模型参数的情况下约束智能体行为。它能在运行时阻止智能体执行未经授权的工具调用或数据操作，解决自主系统可能引发的安全隐患。这为开发生产级AI应用提供了一层轻量级的防护网。

## English Version

**RuntimeGovernance: Deontic Logic Framework for Agentic AI Control**

Addressing security and compliance challenges in LLM-driven agents, a new paper proposes RuntimeGovernance. This framework uses deontic logic to define operational norms, constraining agent behavior without model weight modifications. It blocks unauthorized tool calls or data manipulation at runtime, mitigating risks in autonomous systems. This offers a lightweight guardrail for production-grade AI applications.

---

**来源**：[arXiv cs.AI](https://arxiv.org/abs/2606.19464)

**详情页**：https://ai.daily.yangsir.net/daily/20260620-T0-06

---

*智语观潮 · Daily — https://ai.daily.yangsir.net/llms.txt*