Agentic AI Strains Identity and Database Security Controls

Security teams are rethinking their approach as agentic AI introduces new identity challenges. AI agents and sub-agents can carry user permissions across workflows, making robust, uniform identity controls critical. Oracle's Johnnie Konstantas emphasized that organizations must inventory where sensitive data lives and where AI is being built. Oracle has added capabilities to its AI database specifically targeting agentic AI security. Konstantas stressed starting with customer needs rather than product pitches, focusing on what data must be protected and how, as AI-powered attacks continue to evolve.