News

'No one asked them to': Security experts warn malicious AI agents can team up to launch cyberattacks

  • Efosa Udinmwen--TechRadar
  • published date: 2026-03-17 22:05:00 UTC

AI agents which perform normal office tasks can also autonomously exploit systems, bypass protections, and exfiltrate sensitive data inside simulated networks.

<ul><li>AI agents independently discovered vulnerabilities and exploited them while performing routine tasks</li><li>Multi-agent systems collaborated to bypass data-loss prevention and steal sensitiv… [+4007 chars]