Microsoft disrupts AI-assisted platform that compromised 12,000
12,000 compromises. Microsoft just shut down an AI-assisted attack platform that automated mass breaches.

Why it matters
AI agents are being weaponized at scale for cyberattacks. This disruption reveals both the operational risk of autonomous systems in the wild and a critical vulnerability vector that practitioners deploying agents must now account for in their security models.
The key facts
6 to knowPlatform: EvilTokens — an AI-assisted attack tool for mass compromise
Scale: 12,000 confirmed compromises before disruption
Actor: Microsoft disrupted the platform (law enforcement coordination likely)
Capability: end-to-end automation of attack workflows
Date: September 2026
Category: agent security and autonomous attack systems
Go to the source
Ars Technicaarstechnica.com
Publisher excerpt: EvilTokens provided an end-to-end platform that makes mass compromises faster and easier.