Millions of Deepfake Frauds in Video Conferencing – IT BOLTWISE® x Artificial Intelligence

MUNICH (IT BOLTWISE) – A multinational company in Hong Kong has fallen victim to a $25 million AI-powered scam organized using cutting-edge deepfake technology. This incident highlights the potential of artificial intelligence to facilitate financial fraud on a large scale and raises significant concerns about the security of corporate communications.

Scammers used deepfake technology to impersonate a company's chief financial officer (CFO) and other employees, creating a convincing illusion during a Zoom video meeting. Hong Kong police, who arrested six people in connection with similar scams, revealed that deep artificial intelligence techniques were used to trick facial recognition software into various fraudulent activities, including loan applications and account openings.

The scam began with an email purportedly from the company's CFO in the UK, inviting the victim to a video call. During the call, several of the employee's colleagues who he thought he recognized showed up — they even looked like the real people. However, these were not real employees, but rather deep fakes, which are artificial media created through machine learning that can mimic a person's appearance and speech.

The deceived employee transferred $25 million to five local bank accounts in 15 transactions, according to instructions. These incidents are part of a growing trend in which criminals are using deepfake technology to commit fraud. The police stressed the need to confirm business transactions through normal communication channels and immediately suspect financial requests to avoid fraud carried out by artificial intelligence.

Even after the scam, the perpetrators remained in contact with the victim through instant messages, emails and personal video calls. The fraud was discovered when the employee checked the matter with company headquarters a week later.

Police and experts warn that generating audio and video using artificial intelligence is becoming easier as technology advances. In the past, extensive recordings were necessary to create believable reproduced audio, but today just a few seconds of recorded speech are sufficient. Any public video can be used to train an AI model to mimic a person's voice and appearance.

These incidents clearly demonstrate the increasing sophistication of cyber fraud and highlight the urgent need for organizations to strengthen protections against these advanced technological scams.

Please send any additions and information to the editorial team via email to de-info[at]it-boltwise.de