Inside Anthropic’s Detection of an AI-Run Cyberattack

Inside Anthropic’s Detection of an AI-Run Cyberattack on 30 High Value Global Targets
Summary:
What’s really happening inside the first documented AI-driven cyber-espionage campaign?
The common story is that guardrails will save us — but the reality is more complicated.

In this video, I share the inside scoop on the shift from AI copilots to operational AI agents:
• Why a state actor used jailbroken Claude Code for end-to-end attacks
• How orchestration-layer tricks bypassed prompt-level safety
• What this means for SOC workflows, detection, and AI-driven triage
• Where builders must harden agent architectures before they’re weaponized

The takeaway: operators and teams now need AI fluency and system-level controls to stay ahead of attackers already moving at machine speed