Federal prosecutors in Manhattan have charged three people with stealing more than $129 million from companies around the world using artificial intelligence to impersonate chief executives and top lawyers. The AI CEO impersonation fraud ran for about 19 months, prosecutors say, tricking hospitals, banks, a fintech firm and other businesses into wiring huge sums to accounts the conspirators controlled.

The indictment, unsealed on Thursday, November 5, 2026, names Jacob Soall, 41, Sharon Hazan, 54, and Guy Sheetrit, 42. Each faces seven counts, including wire fraud conspiracy and money laundering conspiracy, carrying up to 20 years in prison, plus a mandatory two years for aggravated identity theft.

Key Points

• A Manhattan federal indictment unsealed on November 5, 2026, charges three defendants with stealing more than $129 million from seven companies between July 2024 and February 2026.
• Prosecutors say the group used AI-generated voices and faces to pose as company CEOs, then as attorneys at major law firms, tricking staff into wiring money to accounts it controlled.
• Named victims include hospital systems in Missouri and California, a Rhode Island bank, a UK fintech firm, a Massachusetts mutual fund company and a Florida pet food company.
• All three were arrested overseas — Soall and Hazan in Spain, Sheetrit in the United Kingdom — and await extradition to the US.
• Each faces seven counts carrying up to 20 years in prison, plus a mandatory two years for aggravated identity theft.

What Happened

The scheme ran from July 2024 through February 2026 and victimised seven companies in the US and the UK. Prosecutors say the defendants found employees authorised to move large sums, then used AI tools to create convincing fake voices — and sometimes faces — of the companies' own chief executives.

Employees received calls and video messages that appeared to come from their boss, ordering urgent, confidential wire transfers. After the first transfer, the conspirators allegedly posed as attorneys at major law firms who confirmed the payments were legitimate, a second layer of deception meant to stop staff checking with anyone else.

The stolen funds were moved through accounts controlled by the conspiracy and converted into cryptocurrency to hide the trail. Investigators are still tracing where the money went.

How the AI CEO Impersonation Fraud Worked

The defendants did not need to break into any computer system. They used publicly available recordings, earnings calls and interviews to train AI voice models that mimicked a chief executive's tone and speaking style well enough to fool staff who spoke to that executive regularly.

In some cases, the group also used AI-generated video faces, so a video call appeared to show the real chief executive giving instructions in real time. With spoofed phone numbers and email addresses, the impersonation was convincing enough that employees followed the instructions without verifying them.

Security experts have long warned that CEO fraud would become far more dangerous once AI could clone voices cheaply. The sums here dwarf earlier cases, which typically relied on simple spoofed emails.

The Victims and the Money Trail

The indictment names six of the seven victims and the sums each lost. A Missouri hospital system lost $4,829,580. A Rhode Island bank lost $2,684,580. A California hospital system lost $2,968,497. A UK fintech company lost $11,578,474, the largest single loss in the charging documents. A Massachusetts mutual fund company lost $4,639,918, and a Florida pet food company lost $5,242,747.

Prosecutors have not named the seventh victim, but say the total across all seven companies tops $129 million. The money was funnelled through conspiracy-controlled accounts and converted to cryptocurrency, which moves fast, crosses borders easily and is hard to claw back once it passes through several wallets.

Hazan allegedly boasted he had laundered money since 1994 and handled more than $400 million in the previous three years. He was previously convicted in Manhattan federal court of conspiracy to commit bank fraud in 2008.

Reactions

US Attorney Jamie McDonald, whose office brought the case, warned that the technology reshaping legitimate business is also being turned against it. "AI offers tremendous potential," McDonald said, but some use it "to pursue criminal goals." A familiar voice on the phone is no longer proof of identity, her office said.

Soall, who prosecutors say called himself "Rocky Balboa," was arrested in Spain in June. Hazan was arrested in Spain in August, and Sheetrit in the United Kingdom in September. All three await extradition to face the charges in Manhattan federal court.

Why It Matters

The case matters far beyond the three defendants: AI has lowered the cost and raised the scale of impersonation fraud. A decade ago, impersonating a chief executive required an insider, an actor or a sophisticated forgery operation. Today, a few minutes of public audio and widely available software can produce a voice clone good enough to deceive employees under time pressure.

Hospitals and banks were among the victims. When a hospital system loses nearly $5 million to fraud, that money is gone from patient care, staffing and equipment. Because the scheme crossed borders — victims in the US and UK, arrests in Spain and Britain, funds converted to cryptocurrency — it shows how hard it is for any single country's police to follow the money alone. For businesses, the lesson is practical: verify any instruction to move money through a separate channel first.

What Happens Next

The three defendants must first be extradited to the United States, a process that can take months if they challenge it in local courts. Once in New York, they will be arraigned in Manhattan federal court, where they are expected to enter pleas.

Prosecutors will keep tracing the stolen funds. Converting the money to cryptocurrency complicates recovery, but investigators have had growing success seizing crypto wallets linked to fraud. If the allegations hold up, the $129 million figure would make this one of the largest AI-enabled frauds ever prosecuted.

Frequently Asked Questions

How much money was stolen?
More than $129 million from seven companies worldwide between July 2024 and February 2026.

Who was charged?
Jacob Soall, 41, Sharon Hazan, 54, and Guy Sheetrit, 42 — each faces seven counts, including wire fraud conspiracy, aggravated identity theft and money laundering conspiracy.

How did the scheme work?
Prosecutors say they used AI-generated voices and faces to pose as company CEOs, then as attorneys at major law firms, tricking employees into wiring large sums to accounts they controlled before converting the money to cryptocurrency.

Which companies were victims?
Named victims include hospital systems in Missouri and California, a Rhode Island bank, a UK fintech firm, a Massachusetts mutual fund company and a Florida pet food company. A seventh victim has not been named.

What sentences could the defendants face?
Up to 20 years per wire fraud and money laundering count, plus a mandatory two extra years for aggravated identity theft.

Related coverage: Chainalysis breach exposes customer names and emails in AI agent attack and DOJ charges green card holders with illegal voting in 5 states ahead of midterms.