By ASPI contributors, External contributors
Publication Date: 2025-11-21 02:35:00
Last week, AI company Anthropic reported with “high confidence” that a Chinese state-sponsored hacking group had weaponized Anthropic’s own AI tools to carry out a largely automated cyberattack on several technology companies and government agencies. According to the company, the September deployment is the first publicly known case in which an AI system performs target reconnaissance with minimal human control.
In a technical report, Anthropic explained how the attackers used their tools to generate code that instructed its agent Claude Code to run the campaign, with human operators responsible for only 10 to 20 percent of the workload. The company did not disclose how it discovered the breach or attributed it to China.
ASPI’s analyzes provide a consistent picture:
—AI reduces the human effort required for sophisticated operations.
– Guardrails designed for individual models will struggle against actors that fragment activities across thousands of…
