Deepfakes have moved from novelty to a practical fraud tool. As synthetic faces and voices get cheaper and more convincing, any identity check that trusts an image or a video at face value is on borrowed time. The response is not panic — it is knowing which checks still hold.
Why a face on a screen proves less
A convincing face is no longer evidence of a real, present person. Printed photos and video replays were already spoof risks; generated faces and face-swaps raise the bar further. Verification that relies on appearance alone inherits all of that risk.
Liveness is the first line
Liveness detection asks the question appearance can not: is this an actual, live person right now? By looking for the signals a flat image or replay can not reproduce, it defeats the common presentation attacks. See liveness detection and anti-spoofing for how it works, and how Live Face fits entry.
Source-verified identity is the anchor
The deeper defence is not trusting the image at all. Verifying a person against an authoritative source — Aadhaar or an issuer-signed document — anchors identity to records a deepfake can not forge, unlike an OCR read of a picture. Combine source verification with liveness and the fraudster has to beat both.
Practical defence at entry
- Verify against a source, do not judge by appearance.
- Add liveness so a credential can not be used with a fake face.
- Escalate low-confidence cases to a person.
- Keep an auditable record of the checks that ran.
Certopact's approach
Certopact Entry is source-verified by design — Aadhaar and DigiLocker with Live Face — so identity rests on authoritative records plus a liveness check, not on an image alone. This is general information, not legal advice.