Tech alert · Scams
Voice-cloning 'grandchild in trouble' calls, in plain English
A short clip of your child's voice, a fake emergency, and a demand for gift cards. The fix takes ten minutes tonight.
What it is
A scammer takes a short clip of a family member's voice, runs it through voice-cloning software, and calls a parent or grandparent with a crying 'grandchild in trouble' — an arrest, a hospital bill, a bad accident. The ask is always urgent and always in a form that can't be traced back: wire transfer, gift cards, crypto, or cash handed to a courier.
The reason it works isn't the technology. It's that panic stops people from making a phone call, and the voice on the phone sounds like the person they love most.
Why children run into it
The voice clips come from public videos. A child's dance video, a school play clip, a sports highlight, a voice message left on a public post — the material is already out there, and the family it belongs to is right there in the comments and captions.
The call usually goes to a grandparent rather than the child, because grandparents are more likely to act fast and less likely to check. The child's posted video becomes the tool used on their own family.
What to do today
Four things, all under ten minutes
- 1
Agree a family safe word tonight
One word or phrase only real family knows. Anyone claiming an emergency who can't produce it is a scammer. This is the FCC's own recommendation and it takes about ninety seconds to settle.
- 2
Hang up and call back on a saved number
Never verify a crisis over the number that called you — caller ID is trivial to fake. End the call, open the contact you already have, and dial. The scam cannot survive that one action.
- 3
Draw a bright line on payment methods
No genuine emergency arrives with a request for gift cards, wire transfer, crypto or cash to a courier. Tell the grandparents in your family specifically, because they are the target and they are the least likely to have heard it.
- 4
Look at what your kids have posted publicly
Not to police it — to see what audio is out there. A public video with a clear voice is the raw material, and it's the one part of this you can change without a argument.
Evidence
What the agencies actually report
The FTC describes the mechanism and the verification steps in a consumer alert. The FBI's Internet Crime Complaint Center reported more than 195 victim complaints and at least $1.9 million in losses for January to September 2023 — a complaint-based count, so a clear undercount of what actually happens, and specific to grandparent scams as a category. Across all internet crime, the 2024 IC3 report logged over $16 billion in reported losses, up 33% on the previous year, with victims aged 60 and over reporting nearly $4.9 billion.
FBI Internet Crime Complaint Center (IC3) · FBI Warns of Scammers Targeting Senior Citizens in Grandparent ScamsWhat we could not establish
- No agency we found states what share of these calls actually use AI voice cloning. The FTC and FCC describe the technique happening; neither publishes a prevalence rate. Anyone quoting a percentage is guessing.
- The $1.9 million and 195-complaint figures cover January to September 2023 only, and count reported complaints about grandparent scams — not AI voice cloning generally, and not unreported incidents.
- We found no official bulletin quantifying losses specifically attributable to cloned voices.
Worth saying
Before you act on this
The safe word only works if everyone knows it
A protocol that lives with one person is not a protocol. Say it out loud at a table where the grandparents are, and write it down somewhere they can find it at 11pm.
This is about your parents as much as your kids
The child's video is the material, but the call goes to someone older. The ten-minute conversation worth having tonight is with your own mother or father.
Settings vs street smarts
Controls set the boundary. Judgement keeps them safe when you're not looking.
A safe word protects the grandparent. It doesn't teach your child to notice when an online 'friend' starts asking them to record a voice message, or to think twice before posting a clip with their voice clear and their name in the caption. That judgement is built by practising the awkward moment somewhere it's safe to get it wrong. NetWise Kids puts them in a realistic group chat, DM or trade request and lets them practise saying no — before the real one arrives.
More alerts
Also worth ten minutes
Evidence
Where every number on this page comes from
Nothing here is estimated. If a figure couldn't be sourced, it was cut.
US Federal Trade Commission, Consumer Advice
20 March 2023
Scammers use AI to enhance their family emergency schemes
Government or regulator
Describes the technique and the verification steps. It does not state how common AI voice cloning is among these calls.
US Federal Communications Commission
Consumer guide, current
'Grandparent' Scams Get More Sophisticated
Government or regulator
Recommends a pre-agreed family safe word and calling the relative back on a saved number.
FBI Internet Crime Complaint Center (IC3)
17 November 2023
FBI Warns of Scammers Targeting Senior Citizens in Grandparent Scams
Government or regulator
195+ complaints and at least $1.9M in losses covers January-September 2023 only, and counts reported complaints — a clear undercount of true incidence.
FBI
23 April 2025
Government or regulator
Over $16B in reported losses across all internet crime, a 33% rise on 2023; victims 60+ reported nearly $4.9B. Aggregate, complaint-based, not specific to voice cloning.