Deepfake

What Deepfake is

Deepfakes have legitimate uses in film, dubbing and accessibility, and harmful uses in fraud, non-consensual imagery and political disinformation. Voice cloning fraud against businesses is now a routine attack.

How it works

Face swapping, voice cloning and full video generation produce the media. Countermeasures include provenance metadata, watermarking, detection models and, most reliably, out-of-band verification procedures for anything involving money or access.

Why it matters

The practical defence for organisations is process, not detection: callback verification and multi-person approval work even when detection fails.

Common uses

• →Threat awareness training
• →Fraud process design
• →Media provenance policy