Today, 20 organizations including the Chinese Academy of Information and Communications Technology, Huawei, Ali, Tencent, Xiaomi, Baidu, and Volcano Engine jointly issued the “Self-Regulatory Convention for Smart Device Development”. Among them, it was mentioned that the intelligent trust system should be improved. Establish hierarchical and verifiable agent identity, and improve mechanisms such as agent authentication and management, authorization and control, and rule expression and embedding. Lower the technical threshold for mutual trust between agents and establish a trust anchor for building an intelligent ecosystem. At the same time, enhance the security protection capabilities of smart devices. Improve technical systems such as forward-looking risk assessment, operation monitoring, behavior control, abnormal fusing, etc., strive to ensure reliable sources of models, plug-ins, and skills, and jointly establish risk information sharing and joint disposal mechanisms. Continuously prevent endogenous artificial intelligence defects and external attacks from security risks such as unauthorized operation, privacy disclosure, and even uncontrolled behavior, and create safe and controllable products and services.

Today, 20 organizations including the Chinese Academy of Information and Communications Technology, Huawei, Ali, Tencent, Xiaomi, Baidu, and Volcano Engine jointly issued the “Self-Regulatory Convention for Smart Device Development”. Among them, it was mentioned that the intelligent trust system should be improved. Establish hierarchical and verifiable agent identity, and improve mechanisms such as agent authentication and management, authorization and control, and rule expression and embedding. Lower the technical threshold for mutual trust between agents and establish a trust anchor for building an intelligent ecosystem. At the same time, enhance the security protection capabilities of smart devices. Improve technical systems such as forward-looking risk assessment, operation monitoring, behavior control, abnormal fusing, etc., strive to ensure reliable sources of models, plug-ins, and skills, and jointly establish risk information sharing and joint disposal mechanisms. Continuously prevent endogenous artificial intelligence defects and external attacks from security risks such as unauthorized operation, privacy disclosure, and even uncontrolled behavior, and create safe and controllable products and services.