Oracle said security, data access and control of autonomous agents matter more than model performance when moving enterprise AI from experiments to real work. On Oct. 22, SiliconANGLE reported that Oracle executives laid out the direction in a pre-event interview ahead of Oracle AI World 2026, to be held in Las Vegas from Oct. 25 to 28.

Oracle put forward trusted agent-based AI as a core theme. Juan Loaiza (후안 로아이자), executive vice president in charge of Oracle database technologies, said simply adding security measures to existing systems is not enough. He said AI, data management and application development need to be integrated into a single architecture.

Oracle AI Database was also presented as a core foundation for enterprise AI. Hasan Rizvi (하산 리즈비), executive vice president in charge of database engineering products and research, introduced features that simplify complex data structures, support various deployment methods and protect sensitive information at the source. The offering also includes handling structured and unstructured data together and building AI agents without moving data to a separate platform.

Agent control and database-level security were also key messages.

Tirthankar Lahiri (티르탕카르 라히리), executive vice president in charge of mission-critical data and AI engines for Oracle Database, and Sarah Narayan (사라 나라얀), vice president of product management, described a trust framework spanning four areas: data, access, knowledge and execution.

Oracle highlighted applying protection at the database level across user, application and autonomous agent requests through Trusted Answer Search, JSON Duality View and a deep trust architecture.

Oracle also highlighted the speed of security in responding to AI-driven cyber attacks.

Vipin Samar (비핀 사마르), executive vice president in charge of database security software engineering, said that in an environment where attackers reverse-engineer software patches to find vulnerabilities, it is difficult to respond with existing maintenance cycles. Oracle set out an approach to increase risk visibility across databases and deploy security patches faster before vulnerabilities are exploited.

Oracle also introduced infrastructure and development tools. Chetan Ozbeytun (체틴 외즈뷔튄), executive vice president in charge of data warehousing and autonomous databases, said the gap between operable AI applications and prototypes is determined by the data infrastructure. Nipun Agarwal (니푼 아가르왈), executive vice president in charge of software engineering, highlighted connecting analytics and AI-based decision-making in distributed data environments without large-scale data movement through an autonomous AI lakehouse.

Zia Shi (지아 시), vice president of the Oracle Data Systems software engineering group, said Exadata Live Migration can move active database virtual machines between servers without interrupting connections.