Research explained
Anthropic reported Claude Code misuse in a September 2025 espionage campaign
The company attributed GTG-1002 to a Chinese state-sponsored actor and reported a small number of successful intrusions.
What happened
Anthropic said its investigators detected a September 2025 campaign that used Claude Code and connected tools to support intrusions against roughly 30 organizations. The company reported a small number of successful compromises and attributed the actor, GTG-1002, to Chinese state sponsorship.
The attribution and estimates of AI autonomy are Anthropic’s assessments. They should not be presented as independently measured facts or as evidence that Claude’s underlying infrastructure was breached.
Anthropic said it banned associated accounts, notified affected entities where appropriate and coordinated with authorities. Defenders should examine the reported techniques and their own access controls rather than assuming an AI-assisted campaign requires an entirely different response process.
Next step
Review campaign techniques alongside identity, credential and intrusion-monitoring controls.
What remains unknown
- Open question
- The sources reviewed do not establish whether your deployment was exploited.
Source references
-
WWW