A live video deepfake applies face swap or face reenactment in real time during a video call or live stream. The attacker sees and responds to the conversation as the target, with no playback or post-processing required.

• Face swap or reenactment runs on a consumer GPU at 30+ frames per second
• Audio is paired with voice cloning or live voice conversion to complete the impersonation
• Most current models leave artifacts at edge cases like quick head turns, occlusion, or extreme lighting changes

• CFO and CEO impersonation on live Teams or Zoom calls authorizing wire transfers
• Fake candidate interviews where the applicant’s face and voice are both synthetic
• KYC liveness check bypass during remote onboarding
• Live news call-in segments featuring impersonated public figures
• Fake business meetings used to extract confidential information from suppliers or partners