标准资料状态Standard Material Status

基本信息Basic Information

标准编号Standard Code

T/ISC 0080—2025

标准名称Standard Title

人工智能通用大模型合规管理体系指南Guidelines for Compliance Management for General-Purpose Foundation Models of Artificial Intelligence

发布日期Publication Date

2025-07-18

实施日期Effective Date

2025-08-18

标准阶段Stage

标准发布Released

英文名称English Title

Guidelines for Compliance Management for General-Purpose Foundation Models of Artificial Intelligence

起草人Drafters

张夕夜、陈慧、郭家书、程晓蕾、杨妮、施克游、赵金生、郝克勤、葛思凡、姜丽丽、邢璟、郝乙入、李鹏、夏晓晖、浦洋、金星、应晓冬、庞成军、张紫阳、聂子尧、江志鹏、郭麒、白海建、徐忠旺、高宏玲、翟腾、艾文思、刘亚洲、郭敏、吕律。Zhang Xiye, Chen Hui, Guo Jiashu, Cheng Xiaolei, Yang Ni, Shi Keyou, Zhao Jinsheng, Hao Keqin, Ge Sifan, Jiang Lili, Xing Jing, Hao Yiru, Li Peng, Xia Xiaohui, Pu Yang, Jin Xing, Ying Xiaodong, Pang Chengjun, Zhang Ziyang, Nie Ziyao, Jiang Zhipeng, Guo Qi, Bai Haijian, Xu Zhongwang, Gao Hongling, Zhai Teng, Ai Wensi, Liu Yazhou, Guo Min, Lv Lv

标准范围Scope

本文件规定了在组织内开展、支持和持续改进人工智能通用大模型合规管理工作的要求并提供指导。本文件适用于下列主体：a)利用通用大模型技术向公众提供生成文本、图片、音频、视频等产品或服务；b)正在开发或委托第三方开发通用大模型。本文件适用于下列范围：a)训练/部署大模型的云平台的安全；b)数据处理全生命周期管理；c)模型安全（含备案、评估、内容安全）；d)服务规范。本文件适用于下列情形：a)计划提升通用大模型合规管理水平；b)寻求外部组织对其通用大模型的合规性进行评价。This document specifies and provides guidance on requirements for carrying out, supporting, and continuously improving compliance management for general-purpose large AI models within organizations. It applies to providers using such models to deliver generated text, image, audio, or video products or services to the public, organizations developing or commissioning such models, and scenarios involving cloud-platform security, data lifecycle management, model security, service specifications, compliance improvement, and external compliance evaluation.