On June 29, 2026, the National Standardization Administration of China (SAC) announced a new national standard project, 20263116-Q-252 General security requirements for artificial intelilgent agent application. The project will be supervised by the Cyperspace Adiministration of China (CAC). SAC/TC260 (Cybersecurity) will administer the formulation. China Mobile, China Electronics Standardization Institute (CESI), and the National Computer Network Emergency Response Technical Team/Coordination Center of China (CNCERT/CC) will lead the drafting. The project has an 18-month cycle and an English version will be developed in parallel.

The standard targets public-facing AI agent products and services. Covered applications may be deployed on mobile phones, tablets, computers, wearable devices or cloud platforms, and may call system functions or local tools to complete user tasks. It applies to producers, developers and other entities throughout design, development, distribution and deployment, and operation.

The main technical content includes security requirements related to:

• Identity identification,

• System permission invocation,

• Tool invocation,

• Data collection and usage,

• Human intervention for high-risk operations,

• Input and output security protection,

• Log retention and dynamic monitoring, as well as

• Anomaly blocking and emergency shutdown.

The project is expected to translate high-level obligations under the Cybersecurity Law, Data Security Law, Personal Information Protection Law, and the Interim Measures for the Administration of Generative Artificial Intelligence Services into technical requirements that can be implemented, tested and assessed.

The drafting process may also draw on existing work by SAC/TC260 (Cybersecurity), MIIT/TC1 (Artificial Intelligence), the China Communications Standards Association (CCSA), and relevant activities at international organizations such as the International Telecommunication Union (ITU). MIIT/TC1 is the ministry-level technical committee responsible for AI-related sector standards in China. Given that AI agents operate across tasks, systems and external tools, the standard is likely to follow a lifecycle approach rather than focusing on a single development or deployment stage.

If adopted as planned, the standard may provide a technical basis for future supervision, market access and technical evaluation. For companies, compliance work may move from policy review into product architecture. Foreign firms participating in relevant technical committees in China are therefore advised to actively engage in the standard’s formulation and related meetings, so they can track its development and identify early signals that may affect product planning and feature design.

If you have any questions or need further assistance, please reach us at: info@bestao-consulting.com.