AI Security · AI-Powered Attack

Generative video turns liveness checks into guessing

Selfie-video verification now has to distinguish a real person from a synthetic one that behaves like one. Static-photo spoofing and simple replay attacks are no longer the main problem when a generated face can move naturally enough to satisfy proof-of-life checks.

Synacktiv says it used generative video to bypass an age-verification system on a site relying on AWS Rekognition, after surveying current verification methods and countermeasures. The issue reaches beyond that product: any KYC, age-gate, or account-recovery flow that treats webcam liveness as proof of a present human is exposed to the same kind of synthetic-video impersonation.

For identity and trust teams, the practical shift is that the control can fail even when the video looks temporally consistent. The question is no longer only whether a face is real, but whether the entire stream is synthetic enough to pass as live.

1 source · Jul 6

Timeline

Sources

Part of the PlainSec briefing for 2026-07-06

Every edition of this story: Generative video turns liveness checks into guessing

More from today