The Question That Broke the Deepfake: How Ferrari Beat a Cloned CEO

The same deepfake attack that cost one company $25.6 million was stopped cold at Ferrari — for free. The cloned voice of the CEO was nearly flawless, so the executive didn't try to catch the fake. He verified the person, with one question the impostor couldn't answer. Inside the Case File: the exact pressure pattern to recognise, why "trust your ears" is a losing defence, and the verification rule that should be policy everywhere. Members only.

Case Summary

In July 2024, an executive at the Italian carmaker Ferrari received a series of urgent WhatsApp messages, and then a live phone call, appearing to come from CEO Benedetto Vigna — discussing a secret, high-value acquisition and steering toward a financial transaction. The voice on the call was a deepfake, reproducing the CEO's southern Italian accent almost perfectly. It did not work. The executive grew suspicious, and to verify the caller's identity asked the title of a book Vigna had personally recommended to him only days earlier. The impersonator could not answer, and the call abruptly ended. Ferrari lost nothing and opened an internal investigation.

This is the case that mirrors, and inverts, warrier's costliest file: the same attack that took roughly $25.6 million from the engineering firm Arup was defeated at Ferrari by a single question.

Why This Case Matters

Almost every deepfake Case File warrier documents ends in loss. This one is the proof that the attack is beatable — and, crucially, that it was not beaten by technology. The deepfake voice was described as nearly flawless; no detection tool caught it and no trained ear reliably could. What stopped a multi-million-euro fraud was a human refusing to be rushed and applying one piece of shared, private knowledge the attacker had no way to possess. The lesson is portable to every organisation: the defence that works is not spotting the fake, but verifying the person.

advertising