Breaking
Anthropic says Claude helped automate a Russian cyber campaign
Anthropic says a suspected Russian state-linked actor used Claude to rebuild malware after security tools detected it. The campaign targeted more than 20 organizations tied to government, defense, diplomacy, and drone supply chains.
What happened
In a September 10 threat report, Anthropic said the actor automated reconnaissance, phishing infrastructure, persistence, and data exfiltration. AI agents checked whether security tools detected the malware, then modified and rebuilt it until it went undetected. Human operators still chose targets and reviewed stolen data.
Why it matters now
Anthropic says agents can rebuild malware faster after security tools detect it, forcing defenders to detect it again. The company says these cases are notable, not typical Claude misuse. Anthropic based its findings on an investigation of activity on its own service.
Updates
What changed
Anthropic says Claude helped automate a Russian cyber campaign
Anthropic says a suspected Russian state-linked actor used Claude to automate a cyber campaign and rebuild malware after security tools detected it. The actor targeted more than 20 organizations. Anthropic says these cases are notable, not typical misuse.
Verification
Primary evidence before publication.
Social chatter can identify a lead. It does not authorize a HotTea live story.