Oracle to require biometric approval for AI agent actions

Oracle plans to extend biometric identity assurance to sensitive operations initiated by AI agents, requiring verification from an authorized employee before proceeding. It also launched AI Security Mesh, which uses OpenAI cybersecurity models to find vulnerabilities, investigate exploits and test fixes.

• OCI biometrics include selfie, liveness detection and ID document checks
• AI Security Mesh uses OpenAI models to probe and test fixes
• Each agent runs under a separate identity with just-in-time access
• Remediation changes require customer approval before applying