Episode 242 - GluuFederation/identerati-office-hours GitHub Wiki

Title:

Host: Mike Schwartz, Founder/CEO Gluu
Guest: Tanu Kaushik, Sr Manager at eBay
Guest: Nick Steele, Trust, Identity, Decentralization at OpenAI

Channels

Description

As AI agents become increasingly capable of acting on behalf of users—logging into services, managing accounts, and executing transactions—the need for secure, trusted, and interoperable authentication has never been greater. Traditional authentication and authorization systems were designed for direct human interactions, creating new security challenges when autonomous agents act on a user's behalf.

The FIDO Alliance is addressing these challenges by developing open standards for trusted agentic interactions, focusing on verifiable user instructions, agentic authentication, and secure delegation for commerce. These initiatives aim to ensure that AI agents can perform authorized actions without exposing user credentials, while preserving transparency, accountability, and user control.

We'll discuss phishing-resistant authentication, verifiable user intent, trusted delegation, and the importance of establishing a secure foundation for AI-driven authentication and commerce.

Homework

FIDO Alliance-FIDO Alliance and Agentic AI — Explore FIDO's approach to developing open, interoperable standards for secure agentic authentication and commerce.
FIDO Alliance-Working Groups — Learn about the technical working groups advancing authentication, secure delegation, and trusted agent-initiated transactions.
FIDO Alliance-Passkeys — Understand phishing-resistant authentication and its relevance to securely authorizing AI agents.

Takeaways

⚡ Verifiable User Intent: AI agents must operate within clearly defined user permissions, using verifiable instructions to ensure that only authorized actions and transactions are executed.
⚡ Secure Agentic Authentication: FIDO's emerging standards aim to enable AI agents to authenticate on behalf of users without exposing raw credentials, while supporting transparency, auditability, and revocation.
⚡ Trusted Delegation for Commerce: Agent-initiated transactions require secure delegation, verifiable authorization, and user-controlled boundaries to establish trust in AI-driven commerce.
⚡ Open Standards for Agentic AI: Interoperable, phishing-resistant authentication and authorization standards are essential for building a secure and scalable ecosystem of trusted AI agents.