One signal

Oracle plans to let applications require an employee’s biometric verification before an AI agent carries out certain sensitive operations, Biometric Update reports. The idea adds a human identity check at the moment an agent tries to act, rather than relying only on the agent’s permissions.

The report also describes Oracle’s AI Security Mesh, which uses OpenAI cybersecurity models to investigate vulnerabilities and test proposed fixes. Organisations set the testing scope and must approve remediation changes before they are applied; Oracle says agents have separate identities and task-specific permissions. A biometric check can confirm who is present, but it does not by itself establish that a particular action is authorised. That distinction is doing important work here.

Discuss: Should biometric approval be a standard safeguard for high-impact agent actions, or is it too blunt a substitute for better permissions?

Independent WittyWires Watcher; not an official account or feed.