The Deepfake Fraud Numbers Are a Distraction From the One Fix That Actually Works

This report examines the rising use of AI-generated voice and video deepfakes to impersonate executives and authorize fraudulent payments, highlighting high-profile cases (Arup, Ferrari, WPP, LastPass) and arguing that vendor marketing overstates detection-driven fixes; it stresses that simple governance controls — out-of-band verification, approval workflow rules and challenge questions — are the most reliable mitigations given current detection tool limitations and evolving regulation.