Synthetic Camera & Voice: Spotting Faked Interview Feeds
The most direct way to misrepresent who's in an interview is to fake the feed itself. Virtual cameras can present a pre-recorded clip or a different person's face as if it were a live webcam, and voice-altering software can reshape audio in real time. The interview app treats both as ordinary devices.
How a synthetic feed is assembled
A virtual camera is software that registers itself with the operating system as a webcam. Anything it outputs — a loop, a filtered face, a swapped identity — appears in the camera picker alongside the real hardware. Voice tools work the same way as a virtual microphone. The interview platform can't tell the difference, because to the platform there's nothing to tell: it's just another input device.
Detecting the pipeline, not the pixels
Trying to judge authenticity from the video pixels is a losing race against ever-better generators. The reliable tell is upstream of the picture: the presence of a virtual camera or virtual audio device, the driver signature behind it, and whether the device appeared only for this session. You don't have to prove a face is fake if you can show the feed is coming from a synthetic device.
Signals worth watching
- Virtual capture devices — cameras and microphones backed by software rather than hardware.
- Driver provenance — known virtual-device drivers loaded during the interview.
- Session-only devices — a camera that appears at session start and vanishes after.
- Corroboration — pair device signals with environment context for a confident call.
Key takeaways
- Virtual cameras and microphones look like ordinary devices to the interview app.
- Don't fight the pixels — detect the synthetic device feeding them.
- A virtual capture device appearing only for the interview is a strong signal.
- Combine with other signals for a defensible conclusion.
Verify it's really them
InterviewWatch detects virtual camera and voice pipelines behind a faked feed — before the offer goes out.