Risk analysis and control of large artificial intelligence models
Shilong Li, Li Zhang, M. H. Gu, Jian Chan, Qi Wu · 2024
With the rapid development of artificial intelligence(AI) technology, large AI models have demonstrated immense potential and value in numerous fields. However, the security risks associated with large AI models have also increasingly garnered widespread attention. This paper first outlines the development status and characteristics of large AI models and provides a detailed classification and discussion of the security risks. Next, the paper analyzes the causes of these risks, delving into technical, talent, and policy perspectives. Lastly, the paper proposes countermeasures for the secure development of large AI models, including improving technical regulation, enhancing policy regulations, strengthening talent cultivation, and reinforcing international cooperation, with the aim of providing a theoretical reference for the healthy development of large AI models.