Case 65A1DC · AI security · L2 Practitioner
Deepfake Impersonation Fraud
- 01 What is it?
- 02 How is it abused?
- 03 How do you stop it?
What a strong answer covers
Try it out loud first. Then check yourself:
- Generative audio and video clone a known person's voice or face from short public samples, convincingly enough for live calls.
- Abuse: fake CFO payment requests, helpdesk resets, fake job candidates. Arup lost about US$25m in 2024 to a deepfaked video call.
- Detection tools are unreliable and lag the generators; spotting artefacts cannot be your main control.
- Process controls work: call back on a known number, dual approval for payments, no exceptions for urgency or secrecy.
- Harden identity flows: helpdesk resets need verified identity, not a familiar voice; train staff on these exact pretexts.
If the interviewer pushes back
- Your helpdesk resets MFA after a video call with the employee. How do you redesign that flow for a deepfake world?
- What would you check before hiring a fully remote candidate you have only met on video?