The worlds first AI fraud? 220,000 British executives were cheated when they answered the boss phone

category:Society
 The worlds first AI fraud? 220,000 British executives were cheated when they answered the boss phone


Artificial intelligence software can not only imitate the voice and intonation of others, but also imitate English with German accent. Picture Verdict, UK

If your boss calls you to do something, will you do it indiscriminately? After reading this story, maybe you will have an extra heart.

According to the Daily Mail on September 6, in March this year, a fraudster used AI voice imitation software to impersonate the companys boss, successfully convincing the CEO of a British energy company that he was on the phone with the owner of the German parent company.

On the phone, the fraudster asked the CEO to immediately transfer 220,000 to a bank account of a Hungarian supplier in order to avoid paying a late fee and send the transfer information by mail, which the executive did without hesitation. Ultimately, the 220,000 euros are gone forever...

Euler Hermes, which insured the energy company, recently disclosed details of the case to the media. Reportedly, the case is considered to be the first fraud involving AI in the world, which has once again raised concerns about the use of AI software in crime.

German accent can also be imitated in English

According to the Washington Post, the CEO later pointed out in an e-mail that the request was quite strange, but the German accent of Big Boss was so lifelike that he felt he had no choice but to obey it.

The underwriter of the British energy company, Yuli Anyi, said the executive thought he was talking to his German boss and telegraphed the money within an hour as required to help the company avoid default fines.

The artificial intelligence software used by cheaters can not only imitate the voice and intonation of others, but also imitate English with German accent, the insurance company said.

It is reported that the cheater made a total of three phone calls. The second call was made after the remittance of 220,000. The other party called the executive and told him that the parent company would transfer funds to repay the expenses of the British company.

The case raised concerns about the use of AI software in crime. Pictures from Daily Mail

Later in the day, when the executive received a third call, and asked for a second payment for the dummy German owner, he realized something was wrong. Because the transfer repayment funds promised by the other party were not received, and the phone number showed that they were from Austria, he did not pay the second amount out of suspicion, but directly called the German owner himself, and found himself deceived.

(Fake)Johanneseven asked to speak to me when I was talking to the real Johannes! The CEO said in an e-mail.

However, Yuli Anyi said that by the time the truth was discovered, the first remittance of 220,000 had been diverted to multiple accounts through accounts in Hungary and Mexico, and no suspects had been identified and the money could not be recovered. Ultimately, the insurance company promises to bear the full cost of the loss.

Some researchers say this is the first time that artificial intelligence fraud has been publicly reported in the world.

Deep forgery AI technology can also be used by non-professional criminals

According to the Daily Mail, this AI voice imitation technique works by breaking a persons voice into syllables or sounds and rearranging them to form new sentences. Although the system is still in its early stages of development with many flaws, if the perpetrators pretend to be in cars or busy environments, they can hide the suspicious places.

Smaller start-ups, such as Lyrebird, are working to make these tools more widely available free and unrestricted. For example, Lyrebird allows anyone to create a voice avatar by uploading a one-minute voice clip, which is known as the worlds most authentic artificial voice.

But this AI-synthesized audio and AI-synthesized video, the so-called deep fakes technology, has heightened fears that new technologies are eroding public trust, empowering criminals, making traditional communications business transactions, home phones and even presidential campaigns more vulnerable to computer manipulation.

Criminals will use any tool that allows them to achieve their goals at the lowest cost, said Andrew Grotto, a researcher at Stanford Universitys Center for Cyber Policy and senior director of White House cyber security policy. This technology, which sounded amazing ten years ago, is now available to any creative non-professional criminal, he said.

In response to criticism and concerns, a Lyrebird spokesman wrote in a statement: Imagine if we dont release this technology, others will develop it, and who knows if their ideas will be as sincere as ours.

Companies like Lyrebird, which develops the technology, continue to emphasize its positive uses, saying it can help automated telephone systems become more humane and help dumb people speak again. However, the latest reported incident is bound to arouse concerns about the use of unregulated deep forgery technology by criminals, leading to targeted hacker attacks and cyber crime.

Researchers at Symatec, a network security company, said they found three similar cases in which executives were asked to remit money to another account by artificial intelligence scammers. The company disclosed that losses in one case totaled millions of dollars.

Former US President Barack Obamas fake video was produced using deep forgery technology. According to the Associated Press

Obviously, many large technology companies have realized the harm of this technology. Just on September 5th, Facebook and Microsoft launched a new campaign to crack down on deep forgery videos that look so lifelike. A $10 million fund has been invested to create a system that uses a fake database to detect deep fake products.

The goal of this challenge is to create a technology that allows everyone to better detect when AI is being used to modify videos and mislead audiences, says Mike Schroeder, Facebooks chief technology officer.

Source: Red Star News Responsible Editor: Luo Chongwei_NB12082