Three Indicted In $129 Million AI-Driven Executive Impersonation Fraud

Federal prosecutors in New York have unsealed an indictment charging three foreign nationals with using artificial intelligence tools to impersonate corporate executives and lawyers in a sprawling scheme that allegedly stole about $129 million from companies, including U.S. hospitals and financial firms.^[1]

The indictment in the Southern District of New York names Jacob Soall, Sharon Hazan and Guy Sheetrit on counts of wire fraud, aggravated identity theft and money laundering, alleging they spent more than 18 months scripting fake CEO emails, staging AI-assisted phone and video calls, and then converting fraudulently wired funds into cryptocurrency to hide their origin.^[1] Prosecutors say the three were arrested between June and September in Spain and the United Kingdom, and that U.S. authorities have requested their extradition to face trial in Manhattan.^[1]

According to the charging documents, the group allegedly targeted finance staff and outside counsel at victim organizations with emails that appeared to come from top executives, then followed up with calls and video conferences in which AI models generated realistic voice copies and manipulated images to match those leaders.^[1] Once employees were convinced they were dealing with real bosses or lawyers, prosecutors say, they approved large wire transfers that were quickly broken up and moved through a network of bank accounts and cryptocurrency wallets controlled by the defendants and their associates.^[1]

Court papers describe a pattern in which the suspects allegedly drafted detailed scenarios for the fake communications, including precise language about supposed acquisitions, urgent legal settlements or confidential projects, to justify why payments had to be made rapidly and kept secret from others in the company.^[1] Investigators contend that several health care providers and financial services firms in the U.S. and abroad ultimately transferred tens of millions of dollars before discovering the ruse and alerting authorities.^[1]

The case highlights how generative AI, once seen largely as a productivity tool and creative aid, has rapidly become a weapon for so-called "deepfake" fraud that can defeat traditional corporate security training focused on spotting odd phrasing or spoofed email addresses.^[1] Law enforcement officials and cybersecurity experts have warned that voice and video cloning make it far harder for employees to rely on familiar cues when verifying who is on the other end of a call, and have urged companies to add stronger payment verification rules that do not depend solely on real-time conversations.^[1]

If convicted on all counts, Soall, Hazan and Sheetrit would face lengthy prison sentences and forfeiture of any assets tied to the alleged scheme, though prosecutors did not immediately detail how much of the $129 million they believe can be recovered.^[1] Defense lawyers for the three have not yet publicly responded to the charges, and no trial date will be set until extradition proceedings in Spain and the UK are resolved.^[1]

1. CBS News