Soon we’ll know what agentic AI is best used for, as long as we can control it

The world expects a lot from AI, and agentic AI is its next frontier. A multitude of bots harnessed to serve human needs is the dream. But what can we really expect agentic AI to do? And how do we ensure it’s following the rules it’s been given?

Biometric authentication firm authID recently launched its framework for agentic AI. On the latest episode of the Biometric Update Podcast, authID VP of Operations Jeff Scheidel explains the concept.

“What we do is, when someone goes to launch an AI agent, we produce a custom claim that can be transmitted through OIDC and a variety of other protocols, so that agent is invoked only with its human sponsor’s verifiable biometric credential. And that can be passed down through agent to agent, when you’re talking about, for example, shopping.”

“If at any point someone says, ‘you know what, I do want to verify that one more time,’ we can be federated out to and request an additional biometric authentication from the user. Typically, that’s not going to be necessary. But that credential that goes along with that agent provides that scope for what that agent’s allowed to do, and then also its auditability. Who is the human sponsor accountable for what that agent did after the fact? Again, we’re using that public key, that credential that we express, and we’re actually using a Web 3.0 credential for that. This is where we are going forward. That’s going to be a big, big thing for us.”

Listen now: Spotify, Apple, YouTube, Podbean

Runtime: 00:20:53

Article Topics

AI agents  |  authID  |  Biometric Update Podcast  |  biometrics  |  digital identity  |  identity access management (IAM)  |  verifiable credentials