Adversarial AI vs. Agentic Defense: Public Sector Resilience

As of March 2026, the cyber landscape has shifted from regional tension to a state of kinetic-cyber convergence following the outbreak of war with Iran. Public sector Security Operations Centers (SOCs) are now on the front lines, facing a surge of over 60 coordinated hacktivist groups and state-sponsored actors like APT33 (Peach Sandstorm). These adversaries are no longer relying on manual efforts; they are utilizing Adversarial AI to automate exploit discovery and launch hyper-realistic, multi-turn social engineering campaigns at machine speed.

This session explored how public sector SOCs can move beyond traditional "Passive AI Assistants” tools to embrace Agentic AI—autonomous systems capable of proactive threat hunting and closed-loop incident containment. We examined the current risks to critical infrastructure—including the targeting of cloud-to-OT pathways—and demonstrate how Google Threat Intelligence, powered by Mandiant’s frontline expertise and Gemini AI, provides the global visibility and automated reasoning required to defend the public sector in this new era of automated warfare.