摘要
新一代人工智能的兴起为国家治理注入新动能,为国家监察工作的数字化、智能化指明方向。人工智能赋能国家监察具有以“数据全覆盖”实现“监察全覆盖”、从人力监督到智能监督等现实价值。受限于国家监察的初创性与人工智能的局限性,国家监察智能化呈现阶段性特征。国家监察的智能化发展围绕数据、算法、算力三大核心要素,实现监察“块数据”构建、面向国家监察的语言模型嵌入及算力资源的优化供给。针对人工智能存在的数据隐私、数据安全等技术性风险,可从人工智能政府采购的制度化与规范化、建构监察对象数据隐私与安全保护机制和人工智能监察应用的纠错与救济机制等方面予以防范和化解。
The rise of the new generation of artificial intelligence has injected new momentum into national governance and pointed out the direction for the digitization and intelligence of national supervision work.Empowering national supervision with artificial intelligence has practical value in achieving“full coverage of supervision”through“data coverage”,from human supervision to intelligent supervision.Due to the originality of national supervision and the limitations of artificial intelligence,the current intelligence of national supervision presents phased characteristics.The intelligent development of national supervision should revolve around the three core elements of data,algorithms,and computing power,achieving the construction of“block data”for supervision,embedding language models for national supervision,and optimizing the supply of computing resources.To address the technical risks of data privacy and security in artificial intelligence,measures can be taken to prevent and resolve them from the institutionalization and standardization of government procurement in artificial intelligence,the construction of data privacy and security protection mechanisms for monitoring objects,and error correction and relief mechanisms for artificial intelligence monitoring applications.
作者
喻少如
唐成余
Yu Shaoru;Tang Chengyu
出处
《中共天津市委党校学报》
北大核心
2024年第2期45-54,共10页
Journal of the Party School of Tianjin Committee of the CPC
基金
教育部哲学社会科学重大课题攻关项目“全面依法治国视域下司法行政职能定位及作用发挥问题研究”(批准号20JZD021)
西南政法大学校级科研创新项目“数字赋能纪检监察的内在逻辑与推进路径研究”(批准号2023XZXS-029)的阶段性研究成果。
关键词
人工智能
国家监察
数据安全
算法风险
纪检监察智能化
artificial intelligence
national supervision
data security
algorithm risk
intelligence of disciplinary inspection and supervision
