On October 15, at the opening of the 2026 AI Industry Conference in Jinan, the "AI Security Evaluation Framework for the Industry and Information Sector (2026 Edition)" was officially released. To better balance development and security, the China AI Industry Alliance, with support and guidance from the Network Security Bureau of the Ministry of Industry and Information Technology, studied and compiled the Framework together with relevant units.
The Framework builds a layered, three-dimensional security evaluation system from three dimensions — general security, application security and business security — with matching evaluation methods and implementation requirements covering basic commonalities, new-application security and key business-scenario security.
General security covers data security, model security and infrastructure security. Application security focuses on agent security and embodied-intelligence security, setting evaluation requirements for key dimensions such as permission security, behavior security, memory security and tool-call security. Business security covers industrial and telecom scenarios, with industrial security running through R&D, pilot verification, manufacturing, marketing and operations, and telecom security spanning network planning, operation optimization, cloud resource scheduling and security defense.
Participating units in the release included the China AI Industry Alliance, the MIIT AI Standardization Technical Committee, the China Internet Association, the China Communications Society's network and data security committee, and enterprises such as China Telecom, China Unicom, Volcano Engine, Baidu, Ant, MiniMax, StepFun, 01.AI, Geely, 360 and Anheng.
Source: Qilu Evening News / NetEase
