Scammers now clone a loved one’s voice with AI to fake an emergency call, so agreeing on a family “safe word” can defeat them.

A man with a beard talking on a cell phone

A phone call from a panicked grandchild, begging for money to get out of trouble, has long been a scammer’s tool. Now the voice on the line can be a convincing imitation of the real person, generated by artificial intelligence from a short sample of their speech. As the technology makes these emergency scams more believable, a low-tech countermeasure, a family code word, offers a simple way to unmask the fraud.

How the scam has evolved

The family-emergency scam works by impersonating a loved one in crisis, claiming to have been in an accident, arrested, or otherwise in urgent need of money, and pressing the victim to send funds immediately and secretly. The Federal Trade Commission warned that scammers use artificial intelligence to enhance family-emergency schemes, using voice-cloning tools to make the caller sound like the actual relative.

The technology needs only a small sample of a person’s voice, sometimes taken from social media, a voicemail greeting, or a video, to produce speech that mimics them. That makes the impersonation far more convincing than the older version of the scam, in which a caller simply claimed to be a relative and hoped the victim would not notice the unfamiliar voice. Hearing what sounds like a grandchild’s actual voice, distressed and pleading, can override a person’s caution in the moment.


Free for readers: Scam calls targeting retirees change every week. The free Retirement Shield newsletter flags the ones going around and the one tell that stops each. Sign up free.

The safe-word defense

Because the voice can no longer be trusted on its own, families can agree in advance on a code word or phrase known only to them. If a supposed emergency call comes in, the person can ask for the safe word; a genuine relative will know it, while a scammer, however convincing the cloned voice, will not. This simple step turns the tables, requiring the caller to prove their identity with information the AI cannot supply.

The FTC’s guidance on family-emergency scams recommends verifying the caller’s identity before acting, and a prearranged code word is one clear way to do that. Choosing a word that is memorable to family members but not something a stranger could guess or find online makes it effective. Sharing it with the relatives most likely to be impersonated, and with the older adults most likely to be targeted, prepares everyone in advance.

Other ways to verify

Even without a safe word, a person can defeat these scams by hanging up and calling the relative back directly on a known phone number. A scammer will resist this, insisting the situation is too urgent or that the person must keep the call secret, but those very demands are warning signs. Reaching the real relative, or another family member who can confirm their whereabouts, exposes the fraud.

The scams also share the payment red flags common to fraud generally. A demand for money through wire transfer, gift cards, cryptocurrency, or a payment app, along with insistence on secrecy and urgency, is a hallmark of a scam. The FTC’s broader advice on how to avoid a scam emphasizes that legitimate emergencies do not require untraceable payment or absolute secrecy, and that pressure to act instantly is itself a signal to slow down.

Preparing the family

The most effective protection is a conversation before any call arrives. Agreeing on a safe word, reminding older relatives that a familiar voice can now be faked, and establishing the habit of verifying by calling back on a known number all prepare a family to recognize the scam in the moment. Being cautious about how much voice and video is shared publicly can also reduce the raw material available for cloning.

The reassuring point is that even as AI makes the impersonation more convincing, the defense remains simple and human. A cloned voice cannot supply a private code word or survive a callback to the real person. For a retiree who might one day receive a frantic call in a grandchild’s voice, having agreed on a safe word in advance, and knowing to verify before sending any money, is a straightforward way to defeat a high-tech scam with a low-tech precaution.

Choosing and sharing a safe word

A safe word works only if it is chosen and shared thoughtfully. The best choice is something memorable to family members but not easily guessed or found online, so avoiding a pet’s name, a birth year, or anything posted on social media matters. A random word or phrase unrelated to public information is harder for a scammer to discover or talk their way around.

Sharing it with the right people completes the plan. The relatives most likely to be impersonated, such as children and grandchildren, and the older adults most likely to be targeted should all know the word, and it helps to agree on when it would be used, namely any call involving an emergency and a request for money. Because the whole point is secrecy, the word should be communicated in person or through a secure channel rather than posted or emailed where it could be intercepted.

Layering the defenses

The safe word is powerful, but combining it with other habits makes a family nearly scam-proof against these calls. Hanging up and calling the relative back on a known number verifies the situation independently, and reaching another family member who can confirm the person’s whereabouts exposes the fraud. The payment red flags common to all scams apply here too: a demand for money through wire transfer, gift cards, cryptocurrency, or a payment app, along with insistence on secrecy and urgency, is a hallmark of fraud, as the Federal Trade Commission’s guidance on avoiding scams emphasizes. Even as artificial intelligence makes the impersonation more convincing, the defense remains simple and human: a cloned voice cannot supply a private code word or survive a callback to the real person. For a retiree who might one day receive a frantic call in a grandchild’s voice, an agreed-upon safe word and the habit of verifying before sending money are a straightforward way to defeat a high-tech scam with a low-tech precaution.


Free for readers: Want plain-English help keeping more of your money in retirement? The free Retirement Shield newsletter covers scams, benefits, and money you’re owed, a couple times a week. Subscribe free.

This article was researched and drafted with AI assistance and reviewed against the linked primary sources.

Leave a Reply

Your email address will not be published. Required fields are marked *