A state-backed actor used an AI agent to run cyber ops, with 80–90% handled autonomously.

Compromise an AI agent already inside your environment, and the kill chain disappears. It already has access, permissions, and normal data flows—so activity looks legitimate.

�� How AI agents bypass traditional detection models → https://thehackernews.com/2026/03/the-kill-chain-is-obsolete-when-your-ai.html