Oracle Moves Database Security Below AI Agents

Oracle Corp. now places security controls directly inside its databases. This change targets new risks from AI agents. The company aims to protect data at its source. It does so regardless of who accesses it.

This strategy shifts defenses away from individual applications. Traditional app guards can be bypassed by smart prompts. Oracle wants a consistent enforcement point. It ensures permissions follow the user, not the app.

Database controls block agent bypasses

Vipin Samar, Oracle’s SVP, says encryption alone is not enough. He argues security must live in the data layer. This prevents agents from tricking the front end. The database acts as the final gatekeeper.

A demo showed an indirect prompt bypassing app rules. It exposed salary data without proper checks. When controls moved inside the database, that prompt failed. The system blocked access to protected information.

David Knox, Oracle VP, explains the benefit. It provides a provable way to guard data. The system stays constant regardless of the attack path. This removes reliance on agents to enforce their own security.

Fleet visibility catches hidden gaps

Managing thousands of databases creates complex security challenges. AI agents can move easily between different systems. Organizations need a full 360-degree view. This helps them spot configuration drift early.

Oracle Data Safe offers cloud-based monitoring. Database Security Central provides a unified view. It tracks users, sensitive data, and policies. This helps close gaps before they are exploited.

Internal checks intercept every query

External tools cannot fully see internal execution. Oracle sits inside the database to monitor activity. It intercepts every SQL query as it runs. This allows real-time security checks on all data access.

SiliconANGLE reported on this strategy shift. The approach addresses the unique risks of AI. It moves the burden of security from apps. It places it firmly on the data source.