Latest Posts Latest Episodes Free Tools

They Said It Was You on the Phone. AI Cloned Your Voice From 3 Seconds of Video.

They Said What?

“Nobody Could Fake Being My Family Member on the Phone.”

Fact-checked by Steve Rhode, consumer debt and fraud expert since 1994 • Last reviewed June 26, 2026 • Every claim below links to a primary source.

Most money news tells you what happened. I tell you what to do about it.

Every weekday I read the enforcement actions, filings and fine print the outlets skip, and turn them into the one or two moves that actually improve your position — a rate worth moving for, a fee you can refuse, a deadline to beat before it costs you.

In the latest issue (Sep 15): A number you don’t look at is not an asset

I write Your Money Actually most weekdays — actionable money information you will not find anywhere else, and the small decisions that compound. It is free, I sell nothing, and I take no money from any company I write about.

Read Your Money Actually

The verdict: MYTH. Criminals are cloning voices from as little as three seconds of audio pulled from a social media video, then calling elderly parents and grandparents pretending to be a grandchild in jail or a car accident — and demanding cash before you have time to think. The FBI’s 2025 Internet Crime report documented over $5 million in losses from AI voice-clone distress scams alone. The FTC has warned about this threat explicitly — and the defense that works is not technology. It is a family code word.

Who’s telling you this: I’m Steve Rhode. I’ve been helping people with debt and financial fraud since 1994, I filed personal bankruptcy myself in 1990, and I sell nothing — no products, no leads, no referral fees. That’s exactly why I can tell you the truth the people who profit from your fear and confusion won’t.

Well, Actually…

Here is the part that stopped me when I first really understood how this works. You do not need a Hollywood studio to clone a voice anymore. Researchers at Microsoft demonstrated in 2023 that their VALL-E model could replicate a person’s voice — pitch, cadence, speaking style, the works — from as little as three seconds of audio. Your grandchild posted a birthday video on Instagram. They left a voicemail you played near your phone. That is all it takes.

A scammer scrapes those few seconds from a public social media post, runs it through an AI cloning tool (many of which are free), and calls you. The voice on the other end sounds like your grandson. He says he has been in an accident, or arrested, or he is in the hospital and needs you to wire money right now, before the lawyer leaves, before it is too late. The urgency is the weapon. The voice is just the key that unlocks your trust fast enough to keep you from stopping to think.

The FBI documented this in detail in its 2025 Internet Crime Report: Americans over 60 reported more than $7.7 billion in cybercrime losses last year — a 59 percent increase in losses over 2024, per the FBI’s own report — and AI-related fraud specifically cost seniors over $352 million, with more than 3,100 victims over 60 filing AI-related complaints. Distress scams using voice cloning technology generated over $5 million in confirmed losses. Those numbers undercount the reality, because most victims never report it — many are too embarrassed, and many do not even realize a crime happened.

In August 2025, federal prosecutors in Massachusetts charged 13 people in connection with a grandparent-scam ring that prosecutors say stole more than $5 million from over 400 victims — with an average age of 84. The operation ran from a call center in the Dominican Republic, using scripts and personal information about their victims to make callers sound convincingly like a real family member. The Department of Justice press release describes how “closer” employees would follow up a distress call pretending to be the grandchild’s attorney. At least nine of the 13 were arrested at the time charges were announced. Prosecutors called it “cold-hearted.” I call it a template that AI voice cloning makes even easier to run.

They Said
“I’d know if it wasn’t really my family member on the phone. Nobody could convincingly fake that.”
Myth
The Truth

AI voice-cloning tools can replicate a person’s voice from as little as three seconds of audio — enough for one public social media clip. The FTC has warned that “scammers use AI to enhance their family emergency schemes,” and that even someone who sounds exactly like your grandchild on the phone may be a criminal using a cloned voice. The only reliable defense is to hang up and independently call your family member on a number you already have.

FTC Consumer Alert: Scammers use AI to enhance their family emergency schemes (March 2023)

They Said
“AI voice cloning scams — that’s a problem for celebrities and executives, not regular families.”
Myth
The Truth

The primary targets are ordinary grandparents. The average victim in one major 2025 DOJ case was 84 years old. Seniors over 60 reported over $7.7 billion in total cybercrime losses in 2025, a 59 percent increase in losses over 2024, per the FBI’s own report. AI-powered scams cost this group over $352 million last year, with more than 3,100 people over 60 filing AI-related complaints with the FBI.

FBI Internet Crime Complaint Center (IC3) 2025 Annual Report

Why You Were Told This

Nobody told you this, exactly. It is a comfortable assumption. Recognizing the voices of the people you love feels so fundamental, so biological, that it seems impossible to fake. And until very recently, it largely was. Even a skilled impersonator had tells — the wrong laugh, a pause where there should not be one, a word your daughter would never say.

AI voice cloning eliminated that gap. The tools are cheap, available online, and can produce output that fools family members who hear that voice every day. The assumption that your ear is a reliable lie detector is simply no longer true — and the scammers know it before most families do. They are counting on the gap between when the technology changed and when people found out.

The urgency is the other weapon. You hear what sounds like your grandchild’s voice saying “Grandma, I need help, don’t tell Mom.” The adrenaline response of wanting to protect someone you love shuts down the skeptical part of your brain. Scammers engineer this. They train for it. They make you feel that one minute of hesitation will cost your grandchild their freedom or their life.

What to Actually Do

  • Set a family code word right now — today. Pick a word or short phrase that every family member knows and that has never appeared online. If someone calls in distress and cannot produce the code word, treat the call as a scam regardless of how familiar the voice sounds. This is one of the most widely recommended defenses against this scam — and almost no family has done it yet.
  • Hang up and call back on a number you already have. Not a number the caller gives you. Not a callback you find by Googling. The number already in your phone contacts, or one you look up independently. A real emergency will still be an emergency two minutes from now. The caller who says “there is no time for that” is telling you it is a scam.
  • Ask a question only the real person would know. The FTC specifically suggests this: “What kind of dog do you have?” or “Where did you spend Thanksgiving last year?” A cloned voice can sound right. It cannot know the answer to a private family question.
  • Refuse to act on any payment method that cannot be reversed. Wire transfers, gift cards, cryptocurrency, or cash sent with a rideshare driver (yes, that happens — prosecutors in the 2025 DOJ case allege Uber drivers were dispatched to pick up cash from victims) are the payment methods of scammers. Real attorneys and hospitals do not ask for bail in Walmart gift cards.
  • Tell your family about this now, before a crisis. The scam works because victims are flustered and alone in the moment. A conversation today — “here is our code word, here is what to do if someone claims I am in trouble” — takes five minutes and can prevent a devastating loss.
  • Report it to the FTC at ReportFraud.ftc.gov if you or someone you know was targeted, even if no money was sent. Reports help law enforcement track these operations.

Infographic comparing AI voice clone threat (requires 3 seconds of audio, $352M in senior losses) versus family code word defense (secret phrase, preventative barrier against distress scam losses)

Steve’s Take

I have been in this space since 1994 and I can tell you: the scam that works is always the one that weaponizes love and panic at the same time. This one does both. It makes you hear your grandchild’s voice in fear, and it makes you feel that every second of hesitation is a second you are failing them. That combination is designed to bypass every rational thought you have.

Here is what I want you to take away: the defense is not technology. You do not need an AI detector or a special app. You need two low-tech things — a family code word agreed on in advance, and the discipline to hang up and call back on a number you already know. That is it. The scam has no answer for either of those moves.

If you are reading this and thinking “my parents don’t know about this” — that thought is your action item. Call them today. Set the code word. It is the five-minute conversation that actually protects people.

Frequently Asked Questions

How can someone clone my family member’s voice so quickly?

AI voice-cloning tools analyze the pitch, rhythm, tone, and speaking patterns in even a short audio sample and generate a synthetic voice that matches them. Microsoft researchers demonstrated in 2023 that their model could clone a voice from as little as three seconds of recorded speech. Public social media videos — birthday posts, graduation reels, casual vlogs — provide more than enough raw material.

Someone called saying my grandson was arrested and I need to send bail money right away. Is this real?

Almost certainly not. This is the classic “grandparent scam” pattern, now often enhanced with AI voice cloning. Hang up immediately. Call your grandson directly on a number you already have in your phone — not a number the caller provides. If he is truly in trouble, that call will confirm it. Do not send money until you have independently verified the situation, no matter how urgent the caller makes it sound.

What is a family code word and how do we set one up?

A family code word is a word or short phrase — known only to your immediate family — that anyone claiming to be a family member in an emergency must be able to provide. Choose something that has never appeared online and is not guessable from your social media (not a pet’s name or street address). If a caller cannot give the code word, you treat the call as a scam. Text or call your family to agree on one today before you need it.

What if I already sent money thinking it was my grandchild?

Report it to the FTC at ReportFraud.ftc.gov and file a complaint with the FBI at IC3.gov. If you wired money, contact your bank immediately — some wire transfers can be recalled if caught quickly. If you used gift cards, call the card issuer and report the card numbers as used in fraud. Do not be embarrassed: the 2025 DOJ case showed that average victims are 84 years old and were deceived by a sophisticated, professional operation. This is not a failure of intelligence. It is a crime.

How much money do these AI voice-clone scams steal from seniors?

The FBI’s 2025 Internet Crime Report documented over $352 million in AI-related scam losses reported by people over 60, with over 3,100 victims age 60 and above filing AI-related complaints. Distress scams specifically — the fake-family-member-in-trouble calls — generated over $5 million in confirmed losses that year. These figures undercount the true total because most victims never report it.

Can I tell if a voice is AI-generated just by listening?

Not reliably. The FTC warns that even if a voice “sounds just like” your family member, it may be artificially cloned. AI voice cloning has reached a level of quality where family members who hear that voice every day have been fooled. Unusual pauses, slight flatness in the emotional delivery, or a slightly robotic cadence can be tip-offs — but do not count on your ear to catch it. Count on your code word and call-back procedure instead.

Why do these scammers target grandparents specifically?

Several reasons: older adults are more likely to answer unfamiliar calls, often have savings or home equity they can access quickly, may be less familiar with AI technology, and — most importantly — are more likely to act immediately out of love and panic when they believe a grandchild is in danger. The average victim age in a major 2025 federal case was 84. Scammers know exactly who they are targeting and why.

This is my perspective based on over 30 years working with people in financial trouble. Every situation is different and only you can decide what is right for yours. Take this as informed input, not a prescription — and use it to have the conversation with your family today.

The bottom line: AI can clone your family member’s voice from three seconds of a social media video — well enough to fool you in a panicked moment. The defense is a family code word and a call-back on a number you already have. Neither requires any technology. If someone you know still thinks they would just know if a call was fake, send them this.

Dealing With Debt? Understanding your options is the first step. See how all your debt relief options compare — including ones most sites won’t tell you about. The Find Your Path quiz gives a recommendation based on your actual numbers, and the Scam-O-Meter checks any company’s complaint history before you sign. Federal Reserve research shows bankruptcy filers recover faster than those who don’t file.

author avatar
Steve Rhode The Get Out of Debt Guy | Consumer Debt Expert
Consumer debt expert & investigative writer. Personal bankruptcy survivor (1990). Washington Post award-winning author. Exposing debt scams since 1994.