The voice on the phone sounds exactly like your daughter. She is crying, she has been in an accident, and she needs money wired to her right now. Except it is not your daughter — it is a synthetic voice, cloned from a few seconds of audio lifted from social media, and the caller is a criminal. Deepfake scams, once a laboratory curiosity, have become a practical everyday tool of fraud.
Police forces and consumer groups now report deepfake voice and video being used in family-emergency scams, fake celebrity investment endorsements and fraudulent video calls impersonating company executives. The good news is that while the technology is impressive, the defences against it are mostly old-fashioned. At News Scam, we explain how to protect your household.
How the scams typically work
Most deepfake frauds follow a familiar script with a new twist. The criminal gathers audio or video of the target from public posts, clones the voice or face, then manufactures urgency: a relative in trouble, a boss demanding a confidential transfer, an investment apparently guaranteed by a famous face. The FBI has warned specifically about voice cloning being used in family emergency and virtual kidnapping schemes.
The raw material is everywhere. A short clip from a birthday video, a voice note posted publicly, a company webinar — any of it can supply enough samples for modern cloning tools, which need only seconds of audio to produce a convincing imitation. The technology itself is neutral and has legitimate uses in film, entertainment and accessibility, which is precisely why simply banning it is not a realistic defence for ordinary families.
Video calls raise the stakes, because a familiar face on screen feels like proof. But the fakes still have tells: odd lighting around the face, unnatural blinking, audio slightly out of sync with the lips, and above all behaviour that breaks the person’s normal patterns — such as refusing to continue the conversation on another number or channel.
Defences that actually work
You do not need specialist software to beat these scams. You need agreed habits:
- Agree a family code word that a real relative can say to prove it is them
- Always verify urgent money requests by calling back on a number you already have
- Treat any demand for secrecy or instant payment as a red flag, whoever it appears to be
- Tighten social media privacy settings to limit public audio and video of your family
The single most powerful habit is the pause. Scammers engineer panic precisely because panic disables checking. A two-minute delay to make one verification call defeats almost every variant of this scam — guidance echoed in consumer advice from the US Federal Trade Commission.
It also helps to agree a household rule in advance: money never moves on the strength of a single call, message or video, whoever appears to be asking for it. Put the rule in writing if that helps it stick. Rules agreed in calm moments hold up far better than judgement exercised in frightening ones, which is exactly what the scammers are counting on.
The same principle protects you at work. If a video call with an executive or official produces an unusual instruction — a payment, a password, a confidential file — confirm it through a separate, known channel before anyone acts on it. Two minutes of checking is a trivial price for certainty.
Building the habit before you need it
Talk about these scams at home before anyone receives such a call, especially with older relatives who may be targeted and are less familiar with the technology. Agree the code word, agree that no one in the family will ever demand secret transfers, and agree that checking is never rude. Businesses should apply exactly the same logic: payment approvals verified out-of-band, every single time, whoever seems to be asking.
Deepfakes will keep improving, and detection tools will keep chasing them. But the fraud still depends on one thing: a human being rushed into sending money. Slow the moment down, and the technology loses its power — as we explore further in our latest coverage of online fraud.
