China Academy of Information and Communications Technology releases' Research Report on Artificial Intelligence Security Governance (2025) '

Recently, the China Academy of Information and Communications Technology (CAICT) officially released the "Research Report on Artificial Intelligence Security Governance (2025) - Promoting the Practical Framework of Artificial Intelligence Security Governance Industry". It is pointed out that the artificial intelligence industry faces multiple challenges in technology, application, management, and collaborative governance. One is the expansion of endogenous safety exposure through technological development. The inherent characteristics of the model trigger complex safety and controllability issues, and the endogenous safety situation is becoming increasingly severe. Compared with traditional network security, the asymmetry of artificial intelligence security attack and defense has intensified, and technological security has highlighted a new situation of "easy to attack but difficult to defend". Secondly, the extension of applications has led to new security challenges. The application of models not only brings about external security challenges such as iterative application forms, abuse of open source ecosystems, and software supply chain vulnerabilities, but also triggers secondary risk transmission and amplification at multiple levels of individuals, groups, and society. Thirdly, the construction of organizational management system faces new bottlenecks. The black box nature of artificial intelligence technology, the uncertainty of its applications, and the diversity of its industrial chain are becoming increasingly prominent, posing management challenges for different organizational entities such as AI model development, system deployment, and application operation, as well as those with multiple identities. The fourth point is that the mechanism of diversified co governance and collaboration still needs to be improved and perfected. The current industry generally faces the problem of insufficient collaborative efforts in core governance, with unified standards yet to be formed and collaborative mechanisms still needing improvement.