AI 52%
high
2025-03-01
The Construction of China's Artificial Intelligence Risk Governance System and Theoretical Elaboration Based on Risk Regulation Model: A Case Study of Generative AI
中国人工智能风险治理体系构建与基于风险规制模式的理论阐述:以生成式人工智能为例-曾雄、梁正、张辉
清华大学人工智能国际治理研究院
tsinghua_aiig
This research paper argues that building an AI risk governance system is the core issue of AI legislation and proposes a framework based on risk regulation, involving risk assessment, management, and communication, along with a unified governance body.
Document Text
347 characters
搭建人工智能风险治理体系是人工智能立法的核心问题。基于风险规制,即以风险为核心考虑要素,按照风险类别和级别确定优先治理的风险场景,其实质在于合理分配监管资源、提高风险治理的有效性,这也是构建中国人工智能风险治理体系的重要参照。在基于风险规制模式下,需要采取适当措施。首先,构建一个包括风险评估、风险管理和风险沟通三个环节的框架流程,以实现风险管控的集中化。其次,坚持风险分级分类的实施路径。在风险类型上,建立人工智能风险类型库,并根据新技术的演变情况及时更新;在风险等级上,制定明确的风险评级标准,开发相应的风险评级工具和模型,将风险划分为不可接受风险、高风险、中风险和低风险,并为不同等级的风险匹配相应程度的治理措施。最后,设立统一的人工智能风险治理机构,以实现监管权的集中统一。
<br>
<div
Topics
artificial intelligence
risk governance
regulation
Metadata
| Publisher | 清华大学人工智能国际治理研究院 |
| Site | tsinghua_aiig |
| Date | 2025-03-01 |
| Category | report |
| Policy Area | 人工智能治理 |
Verification