Log in Download Integrations Articles News Pricing FAQ Contact
Русский Қазақша 中文
Microsoft Shuts Down AI Agent Hacking Platform

Ars Technica · September 23, 2026 · 1Cifer

Microsoft Shuts Down AI Agent Hacking Platform

Microsoft has taken down EvilTokens, a platform that used AI to automate mass account takeovers, exposing roughly 12,000 accounts. The service worked as an all-in-one kit: it tested stolen passwords, bypassed security checks, and handed attackers ready access, with AI speeding up every step.

Previously, running an attack at this scale required a skilled team and time to assemble the right tools. EvilTokens removed that barrier, turning account theft into an almost automatic pipeline available even to people without deep technical knowledge — meaning businesses can now expect more such attacks, not fewer.

Check whether two-factor authentication is enabled everywhere your team logs in — email, CRM, banking portals, and any service running AI agents. It's also worth reviewing active sessions and access tokens, since stolen tokens are exactly what platforms like this exploit. Role-based access and a protected data perimeter are the foundation for AI agents in 1Cifer: every employee only sees the data and processes tied to their position.

Related stories

Claude Helped Researchers Hack OpenAI's SystemsClaude Helped Researchers Hack OpenAI's SystemsHow to tell if your AI platform accounts have been hacked — and why they're hackers' new targetHow to tell if your AI platform accounts have been hacked — and why they're hackers' new targetBiometrics at mobile operators: new subscriber protection rules and what they mean for companiesBiometrics at mobile operators: new subscriber protection rules and what they mean for companies

Reading us regularly? Add 1Cifer to your preferred sources in Google — our stories will show up in your news feed more often.

Add in Google

All news →