"It's an Order from the CEO and Lawyer": $129 Million Stolen... How Can One Laptop and 1 Minute of Audio Make It Possible?

Duped by AI-Generated Voices, Companies Lose $170 Billion
Three Men Indicted in U.S. for Targeting Businesses
"One Laptop Can Clone a Voice in One Minute"

Three men have been indicted in the United States on charges of using generative artificial intelligence (AI) to clone the voices of corporate CEOs and lawyers at major law firms and defraud multiple companies of $129 million (approximately KRW 175.9 billion).

The New York Times reported on November 8, citing an indictment made public by the U.S. Attorney's Office for the Southern District of New York on November 5, that the scheme had continued for more than a year and a half since July 2024.

Investigators found that the group used AI voice-cloning technology to convincingly imitate corporate executives, then called accounting and finance employees to demand large wire transfers. The criminal group targeted companies in the United States and Europe. First, they approached accounting and finance employees by email, making the messages appear to have been sent by the companies' CEOs. The emails said that a confidential financial transaction was underway.

The group then called the employees using AI-cloned CEO voices and instructed them to immediately transfer funds for the secret transaction. The voices sounded very much like the bosses the employees knew. The perpetrators told employees that the transactions had to be handled urgently and demanded immediate wire transfers.

The fraud did not end there. An accomplice posing as a lawyer at a major law firm contacted the targeted employees again and acted as an attorney to complete the transfer process. In some instances, the group also reportedly attempted video calls using deepfakes modeled on the CEOs' appearances.

In one notable case, an accounting employee at a hospital chain in Missouri received a call from someone whose voice was identical to that of a CEO the employee knew. The employee trusted the instructions and sent the money, losing a total of $4.8 million (approximately KRW 6.5 billion). A UK fintech company, which lost $11.6 million, a pet food company in Florida, which lost $5.2 million, and a mutual fund manager in Massachusetts, which lost $4.6 million, were also targeted using the same scheme.

The group created and used a handwritten call script titled "Lawyer Call." It included regulatory terminology and information about insider trading risks. The aim was to emphasize to employees who answered the calls that the transactions involved sensitive legal and financial matters and must not be disclosed to outsiders.

The defendants are Israeli nationals Sharon Hazan, 54, and Guy Sitrit, 42, and Jacob Sol, 41, a UK resident. They were arrested in Spain and the UK, respectively, between June and September 2026 and are awaiting extradition to the United States. Investigators found that the group converted the money stolen from companies into cryptocurrency before distributing it. Hazan reportedly even boasted that he had laundered more than $400 million (approximately KRW 545.5 billion) over the past three years.

They face charges including wire fraud, aggravated identity theft and money laundering. Wire fraud and money laundering each carry a maximum sentence of 20 years in prison per count. Aggravated identity theft carries a separate sentence.

Last year, the FBI received more than 22,000 reports of AI-related fraud, with losses exceeding $890 million (approximately KRW 1.2 trillion). Hany Farid, a deepfake expert and professor at Dartmouth University, explained that a person's voice can be cloned using less than a minute of audio and just one modern laptop. He also warned that a "deepfake avatar" modeled on someone can be created from a single photograph.

"People's ability to tell fake audio and video from the real thing is only slightly better than random guessing," he said. He also recounted how a CEO recently told him, "From now on, I think I'll have to handle all work in person." AI-generated voices and videos have become so difficult to distinguish from real people that they are threatening even companies' internal communication and approval systems.

This content was produced with the assistance of AI translation services.