Family Emergency Voice Scams: What to Do and How to Protect Your Family

What to do after an AI voice family emergency scam call

Act in the first hour. Speed matters most if money has left your account, and each step below is something you can do from your phone. If you did not send anything, skip to the reporting and warning steps; they still matter, because scammers often try other relatives next.

• Confirm your family member is safe by calling them on the number you already have. If you cannot reach them, call another relative or a friend, even if the caller told you to keep it secret.
• If you sent money, call your bank or payment provider immediately using the number on your card, say it was a scam, and ask them to stop or recall the payment. For gift cards, contact the card company with the card numbers and receipt.
• If you shared a password, PIN or one-time code, change it now and sign out of other devices.
• Write down the caller's number, the time, what they said and any names they used, and keep voicemails or call recordings in their original form.
• Report it. In the US, report to the FTC at ReportFraud.ftc.gov, to the FBI at ic3.gov, and to your state attorney general.
• Warn the person whose voice was copied and the rest of the family, and agree a code word for next time.

How AI voice family emergency scams work

The caller sounds like a grandchild, son, daughter or partner in trouble: arrested after a car accident, stranded abroad, hurt, or held by someone demanding money. Often a second voice takes over, claiming to be a lawyer, police officer or doctor, which lets the scammer speak in their own voice after the cloned opening. The FTC's guidance describes this pattern and names the warning signs: urgency, secrecy and emotional pressure.

The voice is cloned from short clips. The FTC warned on 20 March 2023 that scammers can clone a loved one's voice from a short audio clip found online, and the FBI's December 2024 public service announcement says criminals generate short clips of a relative's voice to ask for urgent help or demand a ransom. Details such as names, schools and travel plans come from social media, hacked email accounts or simple guesses.

Payment requests follow a pattern too: wire transfers, cryptocurrency, payment apps or gift cards with the numbers read out over the phone. These methods are chosen because money is hard to recover once sent.

How common are voice clone family scams?

Reliable national counts are scarce because victims often do not know a cloned voice was used. The best recent survey data comes from Starling Bank, which published research on 18 September 2024 showing that 28% of UK adults said they had been targeted by an AI voice cloning scam in the previous year. In the same survey, 46% had never heard of such scams, and 8% said they would send whatever money was requested even if the call seemed strange.

Regulators treat the threat as established. The FTC ran a Voice Cloning Challenge in 2023 and 2024 to encourage technical defences, and the FBI has repeated its warnings about cloned family voices in 2024 and 2025.

How do you protect your family from AI voice cloning scams?

You protect your family by agreeing in advance how you will verify each other, so nobody has to judge a voice under stress. Three habits cover most of the risk.

First, agree a code word or safe phrase. The FBI recommends a secret word or phrase for family members, and Starling Bank advises choosing one nobody else knows and never sharing it by text or email. Second, make callback the rule: any urgent request for money gets a call back to the number you already have, every time. Third, refuse the payment methods scammers prefer. No genuine emergency requires gift cards or cryptocurrency.

Talk through the script with older relatives in particular, because grandparents are the classic target. Practise saying "I'll call you back" and hanging up. Reduce what scammers can collect by making social media accounts private and being careful about posting voice recordings and travel plans publicly.

What banks and telecom providers can do

Families carry most of the burden today, but the organisations in the payment path can interrupt the scam. Banks can add friction to unusual urgent transfers by older customers, ask whether the customer was told to keep the payment secret, and warn about family emergency scams at the moment of payment. Telecom providers can label suspected spoofed numbers and give customers easy ways to report scam calls.

Voice deepfake detection helps where the organisation is on the call, for example when a scammer phones a bank pretending to be the victim's relative or the customer. Neural Defend's NeuroCall, currently in Beta, analyses call audio for synthetic speech and returns a risk signal that the bank's call-handling rules can act on, such as holding a transfer until a callback succeeds. Detection cannot see a scam call made directly to a grandparent's phone, so education and verification habits remain the main defence.