AI is giving attackers a head start, Microsoft warns

Help Net Security coverage highlights AI-enabled cyber threats, including accelerated vulnerability discovery, phishing by China-aligned actor TA419 targeting AI policy experts, insider recruitment, AI-agent access persistence, and attacks against AI systems. The most urgent item is Fortinet FortiMail CVE-2026-104286, an actively exploited unauthenticated path traversal and NULL-byte handling vulnerability with a vendor workaround pending fixes. The feed also reports KillSec ransomware activity involving cloud access and data theft, and Android 17 forensic protections against spyware.