He Wired $25M After a Video Call With His Boss. His Boss Wasn't There.

A finance worker sat down for a video call with the company's chief financial officer. Senior managers were on the screen too. By the end of that call, the worker had wired out twenty-five million dollars. Every single person on that call was fake.

This happened in 2024 at a global engineering firm

This happened in 2024 at a global engineering firm called Arup. Investigators say the faces, the voices, the movements — all of it was generated by artificial intelligence. If you've ever joined a video call and trusted the face on the other end, this story is about you. Because the thing most of us treat as proof — seeing someone's face, hearing their voice — just stopped being proof. So what do you trust now, when a face can be faked in real time?

Let's go back to that twenty-five million dollar call. The employee did everything right by the old rules. They saw their boss. They saw the management team. They acted on what their own eyes told them. And their eyes were lying. That's the part that should stop you cold — there was no hacked password, no broken firewall. The technology worked perfectly. Trust is what failed.

And this isn't a freak one-off. Researchers tracking the fintech world found deepfake fraud incidents jumped roughly eight times over in a single year. The accounting firm Deloitte projects that A.I.-driven fraud in the U.S. could climb to forty billion dollars by 2027 — up from about twelve billion in 2023. That's not a slow creep. That's a wall coming at us fast.

The people pulling this off don't need to be

The people pulling this off don't need to be experts anymore. According to the reporting, attackers can build a convincing fake of your face and voice using A.I. tools anyone can grab online. Where do they get the raw material? Your social media posts. A podcast you were on. A recorded work meeting. If your face and voice exist online, the ingredients for a fake already exist too.

Regulators noticed. In November of 11/01/2024, the U.S. Treasury's financial crimes unit issued its first alert built specifically around deepfake fraud. They gave banks a special code to flag it. When the government creates a new label for a crime, that crime has stopped being theoretical.

So businesses are changing the rules. The research firm Gartner predicts that by 2026, nearly a third of companies will stop trusting face verification on its own. Not because faces stopped mattering — but because a face alone can no longer carry the weight. Companies are layering it now — combining the face with device checks, behavior patterns, and cryptographic keys. For anyone who builds a case for a living — a fraud investigator, an insurance examiner — the old question was "does this face match?" The new first question is "is this real at all?" And for the rest of us, it means the next time a relative video-calls asking for money, the face is no longer enough.