The CEO on the Video Call Was an AI Clone — and It Cost $129 Million
An employee at a Missouri hospital joined a video call with someone who looked and sounded exactly like a company executive. A lawyer dialed in to handle the “confidential” part of a transfer. Every detail a person would normally use to confirm they are talking to the right people — the face, the voice, the presence of a trusted lawyer on the line — was generated by AI. The hospital wired $4.9 million before anyone realized none of it was real.
This is not a warning about a hypothetical scam. On November 6, 2026, the U.S. Department of Justice announced an indictment against three men accused of running exactly this scheme from 2024 to 2026, stealing roughly $129 million from victims that include hospitals, a bank, and a pet food distributor. Every fact below comes from The Record’s reporting on the indictment, published November 6, 2026, drawing on the underlying DOJ and FBI filings.
What the indictment says happened
Prosecutors charged Jacob Soall, 41, a Spanish national; Sharon Hazan, 54, an Israeli national; and Guy Sheetrit, 42, an Israeli national, with wire fraud, aggravated identity theft, and money laundering. Each defendant faces up to 104 years in prison if convicted.
Soall and Hazan were arrested in Spain in the summer of 2026. Sheetrit was detained in the United Kingdom in September 2026. All three face pending extradition to the United States.
The fraud worked in two AI-generated layers. First, the group cloned the voice and on-camera appearance of a real company CEO for a call with an employee who had the authority to move money. Second, to head off any questions, they used AI again to impersonate a lawyer supposedly handling the “confidential” side of the transaction — a pressure tactic that is not new, dressed up in technology that is. Told not to discuss the request with anyone because a lawyer was already handling it, the targeted employee had no one left to ask before wiring the money.
According to the reporting, Soall played the impersonated-lawyer role using prepared scripts. Hazan ran the money-laundering side of the operation; she had a prior 2008 bank fraud conviction and reportedly claimed to have laundered $400 million since 1994. The group routed stolen funds through the United States, Europe, the United Arab Emirates, and China before converting the money into cryptocurrency — a path built to outrun any single country’s ability to freeze it.
Who got hit, and how much
$129 million is the total figure in the indictment. The named victims show the range of targets:
- Missouri hospital — $4.9 million
- U.K. tech company — $11.5 million
- Massachusetts finance firm — $4.6 million
- Florida pet food distributor — $5.2 million
- California hospital — $2.9 million
- Rhode Island bank — $2.6 million
Those six victims total roughly $31.7 million; the remaining balance of the $129 million is attributed to additional victims not individually named in reporting so far. The named list is not a roster of Fortune 500 companies. It is hospitals, a community bank, and a mid-size distributor — organizations with the authority to move large sums of money but, in most cases, no security team watching for AI-generated fraud.
Why hearing your CEO’s voice isn’t proof anymore
For decades, a familiar voice or face on a call was treated as identity verification by default: if it sounded and looked like the boss, it was the boss. That assumption is exactly what this scheme was built to exploit. AI voice and video cloning has become cheap and convincing enough that the checks people instinctively rely on — does the voice sound right, is the video believable — no longer reliably separate a real executive from a well-funded fake.
The second AI layer, the fake lawyer, is what made the scheme work at scale. Cloning a voice is not enough on its own; the target also has to skip the verification step they would normally take. Inventing a “confidential” lawyer who is “already handling it” removes the two or three people an employee would typically loop in before approving a large transfer.
FBI officials point directly at the technology behind the scheme. As quoted in reporting on the indictment, a senior FBI official, James Barnacle, said: “The alleged scheme orchestrated by Soall, Hazan, and Sheetrit demonstrates how the use of advanced technology can be exploited to commit large scale fraud.” That is the plain version of the lesson here — the same technology that makes a video call feel personal and trustworthy is the technology that can fake one.
What to do about it
None of this requires new software or a security hire. It requires a rule that applies no matter who appears to be asking.
- Verify any money-moving request through a channel the requester didn’t choose. If a “CEO,” a “lawyer,” or anyone else calls, emails, or video-calls about a wire transfer, call back on a number already on file before sending anything — never a number given to you during the same conversation.
- Treat “keep this confidential” as a reason to verify, not a reason to skip it. Legitimate deals do not require an employee to bypass the normal approval chain. A request that specifically discourages checking with anyone is the request most worth checking.
- Write the callback rule down as policy, not a judgment call. Anyone with wire authority should follow a short script: thank the caller, hang up, dial the number already in your system, confirm the request. Make it mandatory regardless of rank, urgency, or how convincing the call sounded.
- Don’t let seniority override the check. These scams work specifically because employees don’t feel able to question someone who looks and sounds like the CEO. State explicitly, in writing, that verifying a “CEO request” is expected procedure, not insubordination.
The technology in this case did not need to be flawless to work. It only needed to be convincing enough, paired with urgency and a fake lawyer, to get past an employee’s instinct to double-check. After this indictment, a video call or a voice on the phone is no longer, by itself, proof of who is on the other end.
We break down AI news for business owners in plain language — more explainers in the AIWAS blog, including what a deepfake voice call drill looks like, and short alerts on X at @AIWASai.
Stop leaving customers on read.
We answer your reviews, keep your social pages posting, and make your ads convert, so you get more customers while you run the business.
Get Started