Deepfake Financial Fraud: $40B Problem, 1 Habit to Stop It

A finance worker in Hong Kong joined a video call with his company's chief financial officer and several colleagues. Every single person on that call was fake. By the time anyone found out, he'd sent about twenty-five million dollars across fifteen separate transfers.

If you've ever been on a video call and trusted the

If you've ever been on a video call and trusted the face on the screen, this story is about you. That instinct, seeing someone and believing them, is what these scammers are built to exploit. It's unsettling, and it's fair to feel that way. A new report from a company called identifAI says deepfakes are now tied to a forty-billion-dollar fraud problem. Criminals pull video and audio of real executives from social media, interviews, and podcasts. Then they use A.I. to build a copy that looks and sounds right. So if a face on a screen can't prove who someone is, what can?

Start with that forty-billion figure, because it's easy to misread. It's a projection, not money already lost. Deloitte estimates that fraud powered by generative A.I. will cost the United States about forty billion dollars by twenty twenty-seven. In twenty twenty-three, that number was about twelve billion. So it's a forecast. But it rests on real incidents, and the report says deepfake attacks have risen more than twenty-fold worldwide.

Where do the fakes come from? According to the identifAI report, the raw material is public. Fraudsters gather material on executives from X and LinkedIn. X alone accounts for about forty percent of that research activity. LinkedIn adds a smaller but real share. Your public profile, your conference clip, your podcast guest spot, they can all become training data.

Then comes the method

Then comes the method. The report says video synthesis drives about forty-three percent of deepfake fraud cases. Voice cloning adds roughly twenty-two percent. Together, that's nearly two out of every three cases. So the two things we trust most on a call, a face and a voice, are the two things being faked.

That's not just a corporate problem. The same tools can fake a relative asking for money or a boss asking for a favor.

It also reaches past video calls. Researchers say criminal groups now buy ready-made fraud toolkits. Some use software that fakes a webcam. The system thinks it's seeing a live camera. Really, it's receiving a prepared fake. That lets attackers slip past older identity checks, the kind banks use when you open an account. According to the report, about one in every hundred failed identity checks worldwide now involves a deepfake.