In the past few years, AI techniques have gained wide application in a number of fields, including Image Processing, Speech Recognition, Natural Language Processing, etc. Hence, in security critical application senarios, the security issues of AI techniques have increasingly become the focus of research community and industrial corporations(AI Sec Matrix). Besides their performance, engineers and users should also take the security problems of the AI systems into account, and ensure the satefy of AI models in different business scenarios, avoiding serious consequences induced by malicious control, influence, fraud, faults and privacy disclosure.To provide developers and users a better guidance on the security issues of AI systems, this matrix aims to release a framework to illusrate the attack process and specific attack techniques from the adversaries perspectives, based on ATT&CK paradigm, which is already relatively mature in the network security domain. Understanding and identifying these techniques is helpful for AI developers and maintainers to realize potential risk of AI systems during the overall life cycle and the corresponding solutions, providing essential technical guarantee for the application and deployment of AI systems(AI Sec Matrix).
Artificial intelligence technology has promoted the development of various fields of the economy and society from digitization, informationization to intelligence, but at the same time it is also facing serious security threats. Various threats, from malicious data contamination to the technical vulnerability of the algorithm itself, and privacy leakage, are accompanied by the widespread application of AI technology. Although academia and industry have proposed many defense methods for AI security issues, a systematic solution to AI security issues based on security principles has not yet been proposed. Based on the mature ATT&CK framework in the security field as the theoretical basis for security, we focus on current artificial intelligence risks, covering security issues in the entire life cycle of AI production and operation environments, and propose a set of AI security threat risk matrix to ensure the security of AI models in different business scenarios, so that they will not be easily controlled, influenced, or deceived by attackers, and avoid serious consequences such as misjudgment of results or leakage of privacy data(AI Sec Matrix).