Deepfake

How it works

With a few minutes of public audio or video, attackers can clone a voice or face well enough to pass on a call or in a short clip. They use it to back up a fraudulent request, such as a CFO "confirming" a wire transfer by phone, removing the last bit of doubt a victim might have.

Detection is hard and improving slowly, so process controls like callback verification matter as much as technology.

How it relates to brand impersonation

Deepfakes supercharge executive impersonation by adding a convincing voice or face to a request. The supporting infrastructure, such as fake profiles and lookalike domains, is still detectable and removable.

How nebty helps

nebty cannot stop a voice clone on a call, but it removes the fake profiles and lookalike domains that deepfake-backed scams rely on to look legitimate, and monitors for impersonation of your executives.

How to defend against deepfake fraud

You cannot reliably detect a good deepfake in the moment, so the defence has to be process, not eyesight. The single most effective control is out-of-band verification: any urgent request to move money or release data, even one that arrives by video call in a familiar voice, has to be confirmed through a separate, known channel before it is actioned. Agree a simple internal rule for it in advance, so nobody has to make the call under pressure. Limit how much executive audio and video is available to clone a voice by being deliberate about what gets published. And remember the scam still leans on ordinary infrastructure: the fake profiles, lookalike domains, and spoofed addresses that make the deepfake believable are detectable and removable, even when the synthetic media itself is not.