OpenAI joins FIDO Alliance
OpenAI joined the FIDO Alliance and took a board seat as part of an industry push around AI‑agent authentication and delegated access. (biometricupdate.com) The announcement frames agent authentication and revocation as formal design concerns for agents operating in enterprise environments. (biometricupdate.com)
OpenAI has joined the FIDO Alliance and taken a board seat as the group starts writing authentication rules for artificial intelligence agents. (fidoalliance.org) The FIDO Alliance is the industry group behind passkeys, the cryptographic sign-ins now replacing passwords on phones, browsers, and enterprise systems. FIDO says its mission is to reduce reliance on passwords and set standards for authentication and device attestation. (fidoalliance.org) A passkey lets a person sign in with the same biometric, personal identification number, or pattern they use to unlock a device, instead of typing a password. FIDO says passkeys are tied to a user account and designed to resist phishing and credential reuse. (fidoalliance.org) The new problem is that software agents are starting to act for people inside business systems, making purchases, reading files, and using tools across multiple steps. OpenAI’s own developer documentation says agents can plan, call tools, collaborate across specialists, and keep state while they complete work. (developers.openai.com) FIDO said OpenAI joined to help build “secure and private digital identity frameworks” so agents can be verified, governed by user intent, and trusted in enterprise settings. The announcement also said OpenAI plans to work on authentication for “agentic intelligence.” (fidoalliance.org) That fits OpenAI’s current product push into workplace agents. In February 2026, OpenAI launched Frontier, an enterprise platform for building and managing agents with shared context, permissions, and governance controls. (openai.com) OpenAI’s business site now describes “agent identity & access management” as part of enterprise identity and access management for “employees and AI coworkers.” That language turns a research question into an operations problem for companies that need logs, approvals, and access boundaries. (openai.com) FIDO has been moving in the same direction for months. Its October 2025 Authenticate conference recap said the group was already devoting technical sessions to passkeys, verifiable credentials, biometrics, and agentic artificial intelligence. (fidoalliance.org) The enterprise timing is concrete. FIDO said in February 2025 that 87 percent of surveyed companies had deployed or were rolling out passkeys for employee sign-ins, giving the alliance a large installed base as it extends standards from human logins to software agents. (fidoalliance.org) The hard part is not just proving an agent belongs to a user, but limiting what it can do and cutting that access off when needed. FIDO’s announcement puts verification, user intent, and governance at the center, which is where enterprise buyers usually start when they ask whether an agent should be allowed to act at all. (fidoalliance.org)