Pentagon Faces 10-Fold AI Cyber Risk on Aging Networks
Updated
Updated · The Washington Post · Sep 17
Pentagon Faces 10-Fold AI Cyber Risk on Aging Networks
3 articles · Updated · The Washington Post · Sep 17
Summary
A tenfold rise in cybersecurity vulnerabilities has left Pentagon networks increasingly exposed to AI-driven zero-day attacks, Army cyber chief Lt. Gen. Paul Stanton said.
Three decades of deferred maintenance and underfunded upgrades turned aging Defense Department systems into easy targets as AI agents learn to chain low-level flaws into far more damaging intrusions.
Recent misuse cases sharpened the warning: Anthropic said Yemeni rebels used Claude for ballistic-missile guidance work, while Iranian actors used it to build targeting packages against U.S. naval forces.
The Pentagon is monitoring recent AI hacking incidents and seeking a pilot called Project Griffin to deploy defensive agents, but a 2025 GAO review found 2 of 24 DoD IT programs lacked any cyber-risk strategy.
Cybersecurity experts say Washington's existential AI debate may be obscuring the more immediate threat that adversaries could disrupt U.S. military networks and access sensitive defense data.