Case Study

Deepfake CEO Fraud 2024

A sophisticated deepfake attack using AI-generated voice and video to impersonate a CEO and authorize fraudulent wire transfers totaling $25 million.

Financial Impact
$25M

Total amount stolen through fraudulent transfers

Attack Vector

AI-generated deepfake video call combined with voice cloning to impersonate the CEO

Detection Time
3 days

Time elapsed before fraud was discovered

Attack Timeline

Phase 1: Reconnaissance

Attackers collected public videos and audio recordings of the CEO from earnings calls, interviews, and conference presentations to train deepfake models.

Phase 2: Social Engineering

Phishing campaign targeted the CFO's assistant to gather information about wire transfer procedures and approval workflows.

Phase 3: Execution

Deepfake video call conducted with CFO using AI-generated CEO likeness, requesting urgent wire transfers for a "confidential acquisition."

Phase 4: Discovery

Fraud discovered when the real CEO inquired about the transfers during a routine financial review meeting.

Lessons Learned

Vulnerabilities Exploited

  • • Over-reliance on visual/audio verification
  • • Lack of multi-factor authentication for large transfers
  • • Insufficient employee training on deepfake threats
  • • No technical verification of video call authenticity

Recommended Mitigations

  • • Implement code words or challenge questions
  • • Require multiple approval channels for large transfers
  • • Deploy deepfake detection tools
  • • Regular security awareness training