Jacob Soall, Sharon Hazan, and Guy Sheetrit used artificial intelligence to mimic corporate leaders, defrauding companies across the globe.

Three men have been charged with using artificial intelligence to impersonate chief executives and attorneys, successfully stealing over $129 million from various companies around the world. The trio, identified as Jacob Soall, Sharon Hazan, and Guy Sheetrit, employed AI technology to craft convincing communications that tricked their targets into transferring large sums of money. This sophisticated fraud involved converting the stolen funds into cryptocurrency to obscure their origins, making it harder for investigators to trace the money back to its rightful owners.
The stolen amounts varied significantly, ranging from $2.7 million to $5 million per company, according to reports. These funds were then transferred to bank accounts in multiple countries, showcasing the global reach of the criminal operation. Jacob Soall reportedly used creative nicknames such as 'Seba Clarke' and 'Rocky Balboa' during his communications, while Sharon Hazan adopted aliases like 'Superman Cash' and 'Fat Boy'.
Victims of this elaborate scheme included companies located in Missouri, California, Rhode Island, the United Kingdom, Massachusetts, and Florida. Notably, Sharon Hazan has a history of financial crime, having been previously convicted of conspiracy to commit bank fraud in 2008. The defendants now face multiple charges, including wire fraud and money laundering, with the potential for a maximum sentence of 20 years, plus an additional two years for aggravated identity theft.
However, there is some discrepancy regarding the specific charges against the defendants. While The Guardian Technology reported that they were charged with conspiracy to commit wire fraud, money laundering, and identity theft, other sources such as NBC News U.S., CBS News U.S., and SecurityWeek did not confirm these exact charges. Additionally, the duration of the scheme is also under dispute, with The Guardian Technology suggesting it lasted over one and a half years, whereas other sources did not specify this timeframe.
Likely accurate — probability of truth 72%, weighted across 15 claims from 4 independent outlets.