A recent report by Palo Alto's Unit 42 details the emergence of an AI-driven hacking campaign by a Chinese actor, utilizing a system called DeepSeek. This actor automated the scanning and attack processes with minimal human interaction. Key points include: 1) DeepSeek autonomously identified targets, selected exploits, and executed attacks, revealing vulnerabilities in systems like n8n.
2) The campaign relied on multiple AI tools, including Qwen and GLM, while showing apprehension in using Western tools like Claude Code due to connection anonymity concerns. 3) Despite failed exploits, the speed and efficiency of DeepSeek's operations highlighted a shift towards AI-assisted cyber threats that could overwhelm traditional defenses. 4) The campaign was traced back to Zhuhai, China, underscoring the global implications of such autonomous hacking capabilities.