Deepfake Video Fraud
The use of AI-generated video to impersonate real people in real-time video calls or pre-recorded clips to authorise transactions, extract information, or spread disinformation.
Also known as: deepfake video, AI video impersonation, synthetic video fraud
Last reviewed: 10 June 2026
Deepfake video uses machine-learning models to synthesise facial appearance and movements onto another person's likeness in real time or in recorded video. In fraud contexts, attackers have used deepfakes in video calls to impersonate executives during authorisation requests, in fake video testimonials for investment scams, and to fabricate compromising footage for extortion.
Real-time deepfake video deployed in calls is an emerging threat: an employee on a video call may believe they are speaking to their CEO or a known colleague, when in fact it is an AI-generated avatar controlled by a fraudster. The quality of accessible tools is improving rapidly.
Organisations should establish secondary verification procedures for high-value decisions made over video. Watch for visual anomalies such as blurring at face edges, unusual blinking patterns, or inconsistent lighting.
Examples
- An employee on a multi-person video call is convinced by a deepfake of their CFO to initiate an urgent wire transfer.
- A deepfake video of a public figure endorsing a cryptocurrency is distributed on social media to lure investors to a scam platform.