US Edition
Your source for latest news
USFederal Fraud Case

Trio charged in $129 million scheme that used AI to clone executives' voices

Federal prosecutors say three men used AI-generated voices and video to impersonate corporate executives and lawyers, draining roughly $129 million from companies worldwide.

PW
By PressTemps Washington DeskPublished Today, 09:07 ET · 3 min read
Trio charged in $129 million scheme that used AI to clone executives' voices
The Thurgood Marshall United States Courthouse in Manhattan, home to the Southern District of New York. File photo. Photo: Kidfly182 / Wikimedia Commons, CC BY 4.0
What to know
Three men — Jacob Soall, Sharon Hazan and Guy Sheetrit — were charged by Manhattan federal prosecutors with using AI-cloned voices and video to impersonate CEOs and lawyers, stealing about $129 million from companies worldwide.
Victims named in court papers include hospital systems in Missouri and California, a Rhode Island bank, a U.K. fintech firm, a Massachusetts mutual fund company and a Florida pet food company; all three defendants are foreign nationals awaiting extradition from Spain and the U.K.
The FBI had warned as early as December 2024 that generative AI was making business-email-compromise fraud far more convincing, as reported internet-crime losses climbed to $16.6 billion in 2024.

Federal prosecutors in Manhattan unsealed charges Thursday against three men accused of using artificial intelligence to clone the voices and images of corporate executives and attorneys in a fraud scheme that drained roughly $129 million from companies around the world, according to the U.S. Attorney's Office for the Southern District of New York.

Jacob Soall, 41; Sharon Hazan, 54; and Guy Sheetrit, 42, were charged with conspiracy, wire fraud, money laundering and aggravated identity theft over a scheme prosecutors say ran from July 2024 through February of this year. Soall and Hazan were arrested in Spain earlier this year, and Sheetrit was arrested in the United Kingdom in September; the defendants, who have ties to Spain, Israel, the United Arab Emirates and the United Kingdom, are now awaiting extradition to the United States.

Cloned voices and fake lawyers

Prosecutors said the group targeted accounting and finance employees at companies worldwide with emails that appeared to come from their own chief executive, asking for help with a confidential financial transaction. A phone or video call followed in which artificial intelligence tools mimicked the executive's voice, and at times their image, to make the request seem authentic. Employees were then directed to a supposed outside attorney, also impersonated using AI, who instructed them to wire funds to accounts the conspirators controlled; the money was then converted into cryptocurrency and moved among the defendants, according to prosecutors.

CBS News reported that court papers list specific losses, including nearly $4.8 million from a Missouri hospital system, about $2.7 million from a Rhode Island bank and more than $11.5 million from a United Kingdom financial technology firm. ABC News identified additional victims, including a California hospital system, a Massachusetts mutual fund company and a Florida pet food company, on top of other companies prosecutors did not name. Hazan was previously convicted in 2008 in Manhattan federal court of conspiracy to commit bank fraud, officials said.

The charges carry a maximum sentence of up to 20 years in prison for wire fraud, with aggravated identity theft adding a mandatory consecutive term. U.S. Attorney Jamie McDonald said in a statement that artificial intelligence "offers tremendous potential" for legitimate business, even as the technology is increasingly weaponized to make old-fashioned fraud harder to detect. Lawyers for Soall and Sheetrit did not immediately respond to requests for comment, and it was unclear whether Hazan had retained one.

A warning investigators had already issued

The tactics described in the case track a pattern federal investigators flagged nearly two years earlier. In a December 2024 advisory, the FBI's Internet Crime Complaint Center warned that criminals were increasingly using generative AI to clone voices and stage video calls impersonating executives, making business email compromise schemes far more convincing than earlier versions that often gave themselves away through typos or clumsy phrasing. The bureau's most recent annual tally of internet-enabled fraud found reported losses had climbed to $16.6 billion in 2024, continuing a sharp multiyear rise that officials have tied in part to more sophisticated impersonation techniques.

Business email compromise schemes, in which fraudsters pose as executives or vendors to trick employees into wiring money, have long ranked among the costliest categories of cybercrime the FBI tracks. What investigators describe as new in this case is the layering of synthetic voice and video onto that playbook, letting a small group of alleged conspirators impersonate not just one executive but a chain of supposedly independent figures, including outside lawyers, to walk victims through a transaction step by step.

The case now moves to extradition proceedings in Spain and the United Kingdom before any of the three defendants can be arraigned in federal court in Manhattan. Prosecutors have not said how many companies beyond those named in court papers were targeted, though they have described the victim pool as spanning multiple industries and countries, and said the investigation, led by the FBI, remains open.

More on this story

All US