On April 15, 2026, the 84th working meeting of the China Communications Standards Association (CCSA) TC11 (Mobile Internet Application and Terminal Technology Working Committee) WG3 (Terminal Working Group) came to a successful conclusion in Yichang, Hubei Province. Hosted by Beijing Kuaishou Technology Co., Ltd., the conference brought together many enterprise representatives and technical experts from key links in the industrial chain such as terminal manufacturing, chip design, network security, Internet services and telecom operators. The meeting conducted an in-depth discussion on the industry standard (draft for comments) of "Artificial Intelligence Security Governance and Artificial Intelligence Terminal Security Technical Requirements" formulated by CTTL Terminal Labs of China Academy of Information and Communications Technology (hereinafter referred to as CAICT) , aiming to build industry consensus and build a solid and reliable safety foundation for the healthy and orderly development of the AI terminal industry.

This standard covers key dimensions such as artificial intelligence model security, agent security, operating environment security, and data security, and puts forward technical requirements for the characteristics of AI terminals. Specifically, the standard not only includes the basic security requirements that the terminal needs to meet, but also focuses on the security features of device-side AI, and puts forward clear security requirements for local fine-tuning of AI models, agent security (such as identity authentication, access control, etc.), security isolation of hardware accelerators, protection of the whole life cycle of data, and compliance of generated content.

In the standard discussion and opinion solicitation session, technical experts from Honor, OPPO, vivo, ZTE, Apple, Samsung, Tencent, Douyin, Ant, Hisense Mobile, Haiweitong and other enterprises put forward opinions and suggestions on the specific content of AI terminal security standards based on rich industry experience, mainly focusing on core issues such as term definitions, inclusiveness of technical routes, balance between security requirements and user experience, definition of subjects of different security functions, and synergy with other standards. It aims to ensure that the standard is both forward-looking and leading, and has good implementability, reflecting the openness, pragmatism and collaboration of the standard-setting process.

In the next step, the CTTL Terminal Labs of CAICT will continue to play a leading role in standards, improve the security technical requirements and testing and evaluation system of AI terminals, and work with upstream and downstream enterprises, scientific research institutions and academic experts in the industrial chain to jointly promote the safety, prosperity and innovation of China's artificial intelligence terminal industry ecology.

If you want to inquire about AI terminal and agent security technology and evaluation, please contact us!

Business contact:

Mrs. Yuan

13910330232 yuanqi@caict.ac.cn

Mr. Tian

13126728101 tianchen@caict.ac.cn