Deepfake

FrontierGovernance, Safety and Ethics

A deepfake is synthetic media, including video, audio and images, manipulated or generated using deep neural networks to convincingly impersonate a real person. The business exposure is fraud: social engineering, executive identity theft, unauthorised system access, extortion and brand damage through fabricated communications.

In practice

Deepfakes are a critical threat vector in enterprise identity verification and payment authorisation. Voice approval of a payment is no longer a control. Move high-value authorisation to a channel that cannot be synthesised, and rehearse it, because the attack arrives as an urgent call from a voice the finance team recognises.

Not sure where your organisation stands?

Take the free AI-readiness diagnostic.

Start the diagnostic