AI Voice Cloning Scam: Fake CEO Calls Stole $129M, New York Case
An accountant picks up the phone. It's the CEO, and the voice is unmistakable. The request is urgent and confidential: wire the money now. The CEO never made the call.
What happened?
According to the U.S. Attorney's Office for the Southern District of New York, three men used generative AI to impersonate the voices of chief executives and lawyers. Prosecutors say the scheme cost companies in the United States and Europe more than $129 million.
The same script ran several times over about 18 months, starting in July 2024. The case was reported by The New York Times on November 8, 2026. The allegations are from a federal indictment and have not been tested in court.
How did the AI voice cloning scam work?
The pattern was simple and effective:
- A finance employee gets a call from what sounds like the CEO.
- The caller says the matter is urgent and confidential, and asks for an immediate wire transfer.
- A second voice, a "lawyer", follows up to complete the transaction.
- Both voices are AI-generated.
Prosecutors also say the group tried to pull employees into video calls using deepfake video of the chief executive.
Who lost money?
The victims weren't named, but the amounts were. They include:
- A British financial technology company: $11.6 million
- A Florida pet food company: $5 million
- A Missouri hospital system: $4.8 million
- A Massachusetts mutual fund company: $4.7 million
- A California hospital system: $3 million
- A Rhode Island bank: $2.7 million
Hospitals, a bank, a fund and a food company. No sector was off limits, and none of these victims is tiny.
Why do deepfake CEO fraud attacks work?
They exploit trust, not software. Employees are trained to spot phishing emails with odd links and typos. A familiar voice on the phone bypasses that training. Add urgency and secrecy, and the person on the line has little time to doubt.
Voice used to be proof of identity. It isn't anymore.
How can companies protect themselves?
- Call back on a known number. Never use the number the caller gives you.
- Require two approvers for wire transfers above a set amount, whoever asks.
- Treat urgency plus secrecy as a red flag. Real executives can wait ten minutes for a verification call.
- Verify the media itself. AI audio detection can check a suspicious voice message or recording.
- Protect live calls. Real-time deepfake detection flags manipulated faces and voices during video meetings.
- Brief finance teams. Show them what a cloned voice sounds like before an attacker does.
FAQ
What is an AI voice cloning scam?
A fraud in which criminals use AI to copy someone's voice, often a CEO or lawyer, and use it on calls to trick employees into sending money or data.
How much was stolen in the AI CEO impersonation case?
Federal prosecutors say more than $129 million was stolen from companies in the United States and Europe over roughly 18 months.
Can deepfakes appear on video calls?
Yes. Prosecutors say deepfakes of a chief executive were used in attempts to draw employees into video calls.
How can I tell if a voice is AI-generated?
Human ears are unreliable. Verify through a second channel, and use a detection tool for recordings and live calls.
Don't trust the voice. Verify it.
If a cloned voice can move $129 million, it can move yours. Check suspicious calls, messages and videos before money changes hands.
Try the AI Scam & Deepfake Detector →Source: The New York Times, November 8, 2026. Allegations are from a federal indictment.