Three Men Arrested for $129 Million AI Voice Deepfake Fraud Targeting Global Companies

Overview of the Case

The U.S. Department of Justice's Southern District of New York has prosecuted three foreign nationals for large-scale fraud using AI technology. The suspects are Jacob Soall, Sharon Hazan, and Guy Sheetrit. Soall and Hazan were arrested in Spain, while Sheetrit was arrested in the United Kingdom in September 2024.

The Scam Method: Two-Stage Trust Exploitation

This fraud operates through an exceptionally sophisticated two-stage structure.

Stage One: Impersonating an Executive

• AI technology is used to alter voice and video
• The scammer impersonates a company executive or supervisor via phone or video call
• They request employee assistance with a secret, urgent financial transaction
• When employees see and hear a familiar authority figure, they are unlikely to become suspicious

Stage Two: Impersonating a Lawyer

• The employee is instructed to contact a lawyer at a major international law firm
• That lawyer is also impersonated using AI technology
• A large money transfer is requested
• Since this comes from an independent third party (the lawyer), employees are more likely to trust the instruction

What makes this method particularly dangerous is that it doesn't rely on technical system breaches—it exploits employee trust and psychological vulnerabilities.

Victims and Amounts Stolen

According to the indictment, the suspects stole over $129 million from companies worldwide. Specific examples include:

• Approximately $5 million from a hospital system in Missouri
• Approximately $2.7 million from a bank in Rhode Island
• Approximately $3 million from a hospital system in California
• More than $4.6 million from an investment fund in Massachusetts
• Approximately $5 million from a pet food company in Florida
• More than $11 million from a British technology company

Victims span healthcare, finance, food production, and technology sectors.

Money Flow

Stolen funds were processed through the following steps:

1. Employees transferred money to bank accounts in multiple countries (United States, Spain, Portugal, Greece, United Arab Emirates, Singapore, Austria, Slovakia, China, and others)
2. Funds were then converted to cryptocurrency
3. After laundering to conceal origin and ownership, money was distributed among the suspects and their associates

How to Recognize and Protect Against This Scam

Individual-Level Protections:

• When receiving urgent money transfer requests from your supervisor, independently verify the requester's identity through a separate communication channel
• Understand that even video calls can be deepfaked—don't rely solely on visual/audio confirmation
• Don't use copied phone numbers or email addresses; instead, use officially registered contact information
• Watch for unnatural elements: unusual backgrounds, video glitches, or unfamiliar phrasing

Organizational-Level Protections:

• Require multiple approval steps for all large fund transfers
• Always conduct independent identity verification before transfers
• Provide regular AI fraud awareness training to all employees
• Implement automated systems to detect unusual payment patterns
• Foster a company culture where additional verification of unusual requests is standard practice, even if they come from executives

Official Statement

U.S. authorities emphasize that while AI technology offers tremendous potential, it also poses serious risks of criminal misuse. The suspects face charges including conspiracy to commit wire fraud, conspiracy to commit money laundering, and aggravated identity theft.

Key Lesson

The era of "hearing a familiar voice or seeing a known face means the request is legitimate" is over.

This case demonstrates that employee intuition and trust alone are insufficient. Organizations must implement simple, repeatable verification processes to authenticate trust. As AI technology advances, such fraud schemes will likely increase in frequency and sophistication.