Booz Allen survey shows federal agentic AI adoption outpaces security confidence
BAH•Trust and accountability issues
Trust gaps centered on data protection, unauthorized actions, and prompt-injection resilience; 22% have not assigned accountability for AI-agent incidents.
Threat concerns and readiness gaps
AI-accelerated vulnerability exploitation ranked as the top perceived threat; only 36% are confident defenses can keep pace with AI-enabled attackers.
Readiness to deploy AI-powered cyber defenses was limited, with 31% fully or substantially prepared to integrate them into existing security infrastructure.
Federal AI adoption outpaces security confidence
Booz Allen survey of 105 federal IT and cyber leaders found 58% have deployed or are piloting agentic AI, yet 28% report high confidence in secure deployment.
AI-enabled cyber risk expectations remained elevated, with 79% very or extremely concerned about AI accelerating attacks over the next 12 to 18 months.
