Phishing-Resistant MFA

What Is Phishing-Resistant MFA?

Phishing-resistant MFA is authentication that satisfies what NIST SP 800-63B defines as verifier impersonation resistance.

“The ability of the authentication protocol to prevent the disclosure of authentication secrets and valid authenticator outputs to an impostor verifier without relying on the vigilance of the claimant.” – NIST SP 800-63B

Standard MFA improves on passwords, but most implementations still rely on shared secrets. SMS one-time passwords (OTPs), email codes and push notifications all transfer something an attacker can capture.

An adversary-in-the-middle proxy can collect the OTP and forward it to the real service in real time. The user authenticates; the attacker gets a valid session.

Phishing-resistant MFA eliminates the shared secret entirely. The user holds a private key on a hardware device or secure enclave; the service stores the matching public key. At login, the device signs the server’s challenge, and that signature is scoped to the exact origin. A proxy site receives a signature it cannot use anywhere else.

How Phishing-Resistant MFA Works

Effective phishing-resistant MFA is built around three key properties:

1. Asymmetric cryptography: A new key pair is generated for each service. The private key stays on the authenticator and cannot be exported; the public key registers with the service. Authentication is a signature, not a code.
2. Origin binding: The signed challenge encodes the origin (scheme, host and port). If the user lands on a lookalike phishing domain, the authenticator produces a signature scoped to that domain. The real server’s verification check fails, and access is denied automatically.
3. No human-transferable output: There is no code to read aloud, copy or hand over, so applying social engineering on the user yields nothing usable.

Prompt bombing, where attackers flood users with MFA push requests until one gets approved, appeared in 14% of breach incidents analyzed in the 2025 Verizon Data Breach Investigations Report.

Phishing-resistant methods eliminate this attack entirely because there is no approval prompt to abuse.

Phishing-Resistant MFA vs. Traditional MFA

The following table explains the differences between some commonly-used forms of MFA and evaluates their relative effectiveness.

Types of Phishing-Resistant MFA

There are three main types of phishing-resistant MFA in use today:

• FIDO2/WebAuthn: The FIDO Alliance standard published by the World Wide Web Consortium (W3C). Authentication uses a key pair generated per relying party. Implementations include USB and NFC hardware security keys as well as platform authenticators built into device hardware.
• Passkeys: A passkey is a FIDO2 credential stored in the device’s secure enclave and, optionally, synced across a user’s devices via a platform provider. The FIDO Alliance reported more than 15 billion passkey-capable accounts globally as of December 2024, more than double the count from the prior year.
• PKI-based/PIV/CAC: Personal identity verification (PIV) and common access card (CAC) credentials are X.509 certificates stored on smart cards with hardware-bound private keys. Standard in U.S. federal identity since HSPD-12, they remain the preferred path where hardware security keys are not yet feasible.

Microsoft, under its Secure Future Initiative, reports 92% of its employee productivity accounts are protected by phishing-resistant methods.

Implementation Barriers to Phishing-Resistant MFA

In its fact sheet, “Implementing Phishing-Resistant MFA,” the Cybersecurity and Infrastructure Security Agency (CISA) recognizes common implementation issues and recommends effective solutions:

• Non-compliant systems and vendors: Since not all systems or vendors will support phishing-resistant MFA, start with the most vulnerable supported systems, such as hosted mail and single-sign-on (SSO).
• Large-scale deployments: A single organization-wide rollout may not be realistic. Initially, implement phishing-resistant MFA to your critical systems and teams: IT, security, DevOps, and any other teams that leverage agentic AI.
• Staff pushback: When it comes to new systems and tools, teams don’t always embrace change. For transparency and widespread acceptances, leadership must communicate the benefits of phishing-resistant MFA, including the dangers of credential phishing and the risks of delayed adoption.

Phishing-Resistant MFA: An Effective, but Incomplete Solution

A smart card or passkey may make human logins non-phishable — but AI agents still authenticate with whatever credentials you give them.

Even with MFA, credentials are an extremely vulnerable form of identity verification. That’s why modern enterprises should switch to certificate-based authentication.

“The great thing about certificates is that they are a phishing-resistant form of authentication — not sending credentials over the air,” explains Micah Spady, Product Marketing at SecureW2. “Hackers can’t do anything if they steal something like this. And the cryptography is locked onto a device, so you know if a certificate is shown, it is for sure this compliant device. Now, we can use that information to say, ‘Only certain trusted devices can use my AI systems or pull certain amounts of data in certain AI systems.’”

Phishing-resistant MFA helps prevent credential exposure or misuse by agentic AI, but certificate-based authentication completely eliminates the risk of vulnerable credentials.

Related Terms

Frequently Asked Questions