关键词: AI, 基础设施安全, 网络攻击, 主动防御

Abstract:

Against the backdrop of rapid advancements in Artificial Intelligence (AI), the capabilities of AI-driven cyber-attacks are undergoing a marked leap, profoundly reshaping the global landscape of cyberattack and defense. This paper systematically analyzes the significant enhancements in AI-driven attacks in terms of destructiveness, attack thresholds, and vulnerability discovery capabilities. It reveals the underlying logic that makes critical infrastructure a primary target of AI attacks, and dissects the common threats facing China’s infrastructure from three dimensions: the information layer, the platform layer, and the physical layer. On this basis, the paper further examines the key challenges confronting China’s infrastructure, including security risks arising from intelligent upgrades and the capacity bottlenecks of traditional defense systems. The findings indicate that urgent efforts are required to strengthen security governance during intelligent upgrades and to build a novel defense architecture tailored to the characteristics of AI-driven attacks, thereby reinforcing the security shield for infrastructure. As AI attacks are rapidly evolving from isolated attempts to large-scale threats, the coming years will witness a critical turning point. It is therefore strongly recommended that high priority be given to accelerating the development of a new defense system capable of effectively responding to the distinctive features of AI-enabled threats.

Key words: AI, infrastructure security, cyber-attacks, proactive defense