The National Technical Committee 260 on Cybersecurity of Standardization Administration of China (TC260) has released the Practical Guide to Cybersecurity Standard – Security Emergency Response Guide to Generative Artificial Intelligence Service (the "Guide").
The Guide describes the grading and classification methods for the security incidents in generative artificial intelligence (AI) service, along with the management measures and technical methods to be adopted for the generative AI service security emergency response. The Guide, upon fully considering the causes, threats, attack methods, harms, and other factors relating to generative AI service security incidents, classifies these security incidents into information content security incidents, data security incidents, and cyberattack incidents, among others, pursuant to the classification methods specified in Section 5.1 of GB/T20986-2023. According to the Guide, emergency response process is divided into four stages: emergency preparation, monitoring and early warning, emergency handling, and review and improvement. Among them, detailed management measures for the emergency preparation stage should be implemented in accordance with Section 5.2 of GB/T 20985.1-2017, with the main actions listed.