The Recent Ferrari Deepfake Attempt and the Discipline of the Skeptical Question

The reported attempt to impersonate Ferrari's CEO using a cloned voice is notable not because it succeeded, which it did not, but because of how it was caught. The executive on the receiving end asked the caller a question only the real CEO could answer, the impersonator stalled, and the call was terminated.

That outcome was not a technological win. It was a cultural one. The receiver had been trained, formally or informally, to ask a verification question when something felt off. Most organizations have not made that the default behavior, because it feels rude and slow and often turns out to be unnecessary. The Ferrari incident is a useful counterexample.

Building this discipline is straightforward in principle and difficult in practice. It involves rehearsing the awkward question, normalizing the callback, and giving employees explicit permission to inconvenience executives in the name of verification. It also involves making sure the verification questions themselves cannot be answered from public information.

The attackers in this case were unlucky. The next ones will prepare better. Whether they succeed depends less on the deepfake quality than on whether the receiver has been trained to ask the second question.