A employee of a company in Hong Kong was deceived and pressured into transferring almost 25 million euros of the company during a virtual meeting in which one or more cybercriminals posed as colleagues and managers of the company itself using the deepfake technique.
Second as reported by RTHK (which is a bit like the RAI of Hong Kong), the complaint was apparently presented last January 29, after the woman had been invited to participate in a video conference in which “numerous other people” were expected to take part. But they weren’t actually people.
The guide How to recognize fake news and how to defend yourself in the era of artificial intelligence by Emanuele Capone 27 November 2023
“I’m on call”, but it was the artificial intelligence
According to a spokesperson for the company, whose details were not revealed, the hypothesis is that the scammer has used artificial intelligence to deceive the woman: “He invited her to a video conference in which many other colleagues would have participated” and “seeing that the people present in the call seemed like real people, the employee followed the instructions and carried out 15 money transfers to 5 current accounts which were indicated to her for a total of 200 million Hong Kong dollars.” That is, just under 25 million euros.
According to an initial reconstruction, the scammer would have downloaded videos of people in advance who would be present at the meeting, then it would use AI to reproduce their faces and voices for use during the video conference.
Again according to what RTHK told, the woman would have received a message from the CFO (the financial director) of the company who invited her to maintain maximum confidentiality about the transactions: only after transferring the money and speaking to the company’s headquarters did the woman realize that it was a scam. This confirms that “fraudsters are now capable of use AI in online meetingsso people need to be careful even in those with many participants.”
Google and Universal in negotiations for music produced with artificial intelligence by Bruno Ruffilli 09 August 2023
What are deepfakes
As we have already often explained on Italian Tech, deepfakes are false (fakein fact) videos, images and audio created by applying the principles of deep learning typical of artificial intelligence, which starting from one face is able to simulate another, even recreating facial expressions and expressions. For example, it is on deepfakes that they are based very successful app which allows you to alter your face, see it aged, set in the 1930s, the 1970s and so on.
Simply put, what the software does is access one boundless database of faces of the subject to find the useful pieces to compose what you want to recreate, between eyes, noses, lips, cheekbones and so on, combining them together to arrive at the most realistic and credible result possible.
Deepfakes are a danger not only because what happened in Hong Kong says so, but because it is clear that they will make more and more money difficult to distinguish the real from the fake, news from fake news, the real Donald Trump from the one who ended up arrested in New York and so on. The risk is so great that in recent days Meta has announced its intention to brand content created with AI with a specific label and shared on Facebook, Instagram and Threads and that the European Union is updating its rules on combating the child pornography to also include deepfakes and other material generated with artificial intelligence.
@capoema