The pattern is now depressingly familiar. An employee joins a video call. The CFO is there, colleagues are there, faces and voices are right. Instructions are given, transfers are made, and days later the company learns that every participant on that call except the victim was AI-generated. The most famous case cost a Hong Kong firm roughly 25 million dollars, and 2024 and early 2025 have produced a steady stream of successors across Europe.
Faces are no longer evidence
Video conferencing was never designed as an authentication system, yet businesses treat a familiar face on screen as exactly that. Generative models have broken this assumption permanently. If the approval process for a payment, a contract or a benefits assessment rests on "I saw them on the call", the process is broken.
Gate the room, seal the record
The countermeasure is not better eyesight; it is cryptography at the door. WeVerify's Secure Video Call verifies every participant before they enter: NFC passport chip plus certified biometric liveness, the same checks that catch injected deepfakes. The session is then recorded and cryptographically sealed, producing a court-admissible chain of who was present and what was agreed. The Municipality of Den Haag runs its WMO social-care assessments on this flow with a no imposter accepts recorded to date; application live, rollout underway.
Deepfakes will keep improving. Government-signed chips will not become easier to forge.