Deepfakes have come a long way. From political hoaxes to job interview scams, synthetic video and audio are already being used to trick ID verification systems around the world.

In 2025, we saw reports of North Korean operatives using deepfakes to pass remote job interviews. And according to the World Economic Forum, over 40% of companies faced a deepfake-related threat last year.

But how much of this applies to authentication?

Right now, the biggest impact is on identity verification (IDV), especially video-based onboarding. Fraudsters can replay videos, inject manipulated footage, or spoof interviewers in real-time.

To fool a biometric authentication system at scale, fraudsters would need to:

That’s a high bar. And there are still easier ways to break in, like phishing SMS OTPs or re-registering FaceID on a stolen phone.

But deepfakes are evolving. And in the next 1-2 years, the risk to authentication will grow. That’s why organizations need to prepare now.

Our 2026 report outlines three layers of defense every biometric system, both verification and authentication, must have:

Keyless already deploys all three, including passive liveness detection and certified injection attack resistance.

Deepfakes might not be the main threat to authentication, but in 2026 they will begin to be.

Don’t miss our full analysis in The State of Authentication 2026, launching this October. The State of Authentication 2026 report will cover the five key forces reshaping authentication in the year ahead.