A document with a speech bubble and a row of filled boxes representing a family verification code word.

Grandparent Scams and AI Voice Cloning: What Actually Works

By Tom Reardon

For years, the standard advice on the grandparent scam had a comforting flaw built right into it. The reassurance was always some version of “I would know my own grandchild’s voice.” That used to be true enough to lean on. It isn’t anymore.

The grandparent scam is the call where someone claims a loved one is in trouble and needs money right now. A grandson in a car wreck. A nephew jailed in another state. A daughter stranded overseas with a stolen wallet. The story changes. The pressure doesn’t. And the old defense, trusting your ear, has been quietly taken apart by software that can copy a voice from a few seconds of audio.

I spent over twenty years building identity protection products, and the one thing I want you to take from this piece is simple. The protection that still works is not recognizing the voice. It’s having a habit you fall back on when the voice sounds exactly right and the story sounds exactly wrong.

What This Scam Actually Is

The grandparent scam is one branch of family-emergency imposter fraud. Someone calls pretending to be a relative in crisis, or pretending to be a person helping that relative, a lawyer, a police officer, a doctor, a bail agent. They describe an emergency, and they want money moved before you get a chance to check the story.

It helps to see how this differs from the government impersonation scam, because the defense is different too. A fake IRS or Social Security caller leans on institutional authority. They threaten arrest, a suspended number, a frozen benefit. Skepticism about the agency is your way out. The grandparent scam works the opposite way. It leans on love and obligation. The caller sounds scared, uses a family name, and begs you not to tell anyone. Skepticism feels like betrayal, which is exactly the reaction the script is built to produce.

One thing the official data makes clear is worth saying plainly. There’s no clean national count for grandparent scams on their own. The FBI’s Internet Crime Complaint Center files them under confidence and romance fraud, and the FTC groups family, business, and government impersonation in ways that blur the line. Read any number for this scam as part of a larger imposter-fraud picture, not a precise tally of this one trick.

How AI Voice Cloning Changed the Math

The script above is old. What’s new is how cheap it became to make the call sound real.

The FTC has warned that a scammer needs only a short audio clip, the kind of thing people post online without a second thought, plus a voice-cloning tool, to produce a convincing copy of a relative’s voice. This isn’t a lab hypothetical. In 2025, Consumer Reports tested six voice-cloning products and found four of them let a user create a clone from publicly available audio with no real consent check in the way. The barrier used to be “get hours of clean recordings.” Now it can be “find a short clip and paste it into a website.”

That’s why the “I’d know their voice” defense no longer holds. A 2024 University of Florida study found people could tell real audio from deepfake audio only about 73 percent of the time in controlled conditions. Roughly one in four times, careful listeners got it wrong, and that’s before you add a panicked story, a secrecy demand, and a clock ticking. In the moment this scam creates, your ear is not the reliable instrument you think it is.

Consumer Reports makes a related point worth repeating. Real-time AI detection tools are not a household defense. Nobody convinced their grandchild is sitting in a jail cell is going to calmly run an audio-analysis check mid-call. The fix has to be a habit you already have, not a gadget you reach for under stress.

How the Call Usually Goes

The tools change. The shape of the call has been stable for years. Knowing the sequence is its own protection, because once you can see the next move coming, the spell weakens.

It usually opens with distress. Crying, fast breathing, a panicked “Grandma? It’s me.” Sometimes the caller says almost nothing identifying and waits for you to fill in the blank. The instant you supply a name, they have it.

Then comes the story hook, something easy to picture. A crash, an arrest, a hospital, a wallet stolen on a trip. If the caller already pulled details off social media, the story lands with uncomfortable specificity.

Next is the secrecy command, the single most useful tell in the whole call. The FTC notes that these scammers routinely insist the situation is urgent, that you’re the only one who can help, and that you must keep it quiet. That instruction isn’t an aside. It’s the mechanism. It exists to cut off the one move that beats the scam, which is checking with somebody else.

Often there’s an authority handoff. The crying grandchild passes you to a “lawyer” or a “police officer” who makes the money request sound procedural instead of desperate. Then the payment demand, almost always pointed at rails that move fast and reverse slowly. Wire transfer, gift cards, cryptocurrency, a payment app, sometimes a courier sent to collect cash at the door.

If you hesitate, the urgency escalates. Act now or it gets worse. And if you pay, the danger isn’t over. The same operation may invent a second fee, or circle back weeks later posing as someone who can recover your lost money for an upfront charge. The FTC warns specifically that people who’ve already been hit get targeted again by these fake recovery offers.

Why Urgency Is the Actual Weapon

It’s tempting to think the cloned voice is the dangerous part. It isn’t, not really. The voice is there to rush you toward the part that does the damage, which is speed itself.

Every element of the script pushes you from thinking toward reacting. Cybersecurity research backs up what the scammers already know by instinct. A 2025 University of Washington study on phishing found that stress and time pressure measurably hurt people’s ability to evaluate a threat, and higher stress tracked with worse detection. That study looked at phishing rather than phone calls, so applying it here is an inference, but a fair one. Both attacks run on the same trick, which is shoving the target out of careful thought and into fast habit.

The FTC’s own reporting shows where this bites hardest. In 2024, median losses from phone-based fraud ran higher than fraud that started on social media, around $2,210 versus $650 per report, and adults 80 and older reported notably steeper losses on scams that began with a phone call. A voice-cloned emergency call sits right on that fault line.

The newest figures are thin but pointed. The FBI’s IC3 logged 22,364 AI-related complaints and roughly $893 million in losses for 2025, and it singled out “distress scams” that use voice cloning to mimic a loved one as accounting for over $5 million in reported losses that year. That’s only the slice complainants could tag as AI-linked, so the real number is larger. Either way, this moved from “could happen” to “is happening” some time ago.

The Defense That Actually Works

Here’s the part to keep. Since you can’t trust the voice, you build a process that doesn’t depend on the voice. Three habits do almost all the work.

Hang Up and Call Back on a Number You Already Have

This is the core move, and the FTC and FCC both land on it. Don’t verify the story on the call you’re already on. End it, then call your relative back using the number saved in your phone or written on a list you keep. If they don’t pick up, call a parent, a sibling, a spouse, a roommate, anyone who can physically check on them. Caller ID won’t save you, because a number can be spoofed to show whatever the scammer wants. The only trustworthy line is one you dialed yourself.

Set Up a Family Code Word

Pick a short word or phrase your family agrees on in advance and reserves for real emergencies. When a frightening call comes in, you ask for it. No code word, no money. The trick is choosing something that can’t be guessed or dug up online, so skip pet names, birthdays, schools, and middle names, all of which a scammer may already have. A real secret beats an “interesting fact” every time. This one habit converts the impossible question, “does this sound like them,” into an answerable one, “can they tell me the word.”

Make a No-Fast-Payment Rule

Agree as a family that no emergency money leaves by wire, gift card, crypto, cash courier, or payment app until a second trusted person has confirmed the emergency independently. The scam runs on speed, so a rule that forces a pause is doing exactly the right job. These are also the precise payment methods the FTC ties to fake-emergency fraud, and that’s no accident. They get chosen because they’re hard to claw back.

Tell older relatives that bank friction is on their side.

Say it out loud: it’s fine if their bank slows an unusual transfer. The CFPB encourages setting up a trusted contact with your bank, and some states let financial institutions place a short hold when they suspect an older customer is being exploited. Bank friction isn’t an insult. Sometimes it’s the thing that saves the money.

What to Do During a Suspect Call, and After Money Is Sent

If you’re on a call like this right now, keep it simple.

  • Don’t argue the story on that call. Get off it. You’re not trying to win the conversation, only to leave the scammer’s frame.
  • Call back on a known number. Saved contact first, then another relative who can verify.
  • Ask for the code word if your family has one. Treat a miss as a stop sign, not a debate.
  • Don’t send money by wire, gift card, crypto, cash, courier, or payment app.
  • If the caller claims someone is in physical danger, call 911 yourself or look up the local police number independently. Don’t let the caller be your only source of “official” facts.
  • Save everything. Voicemails, texts, screenshots, numbers, payment instructions, and any names the caller used. It helps your bank and the police later.

If money already went out, the job shifts from prevention to speed.

  • Call the payment channel immediately. Bank, card issuer, wire company, gift-card issuer, the app, the crypto platform. Ask about reversal, freezing, or a refund. Recovery odds vary a lot by method, but fast contact is the consistent advice.
  • If you mailed cash or a package, contact the U.S. Postal Inspection Service right away and ask about intercepting it. If another carrier was used, call them.
  • If you shared personal information, start the identity-theft process at IdentityTheft.gov and change any exposed passwords now.
  • If the victim is an older adult, bring in support early. Adult Protective Services handles suspected elder financial abuse, and the DOJ’s National Elder Fraud Hotline can walk families through next steps.
  • Expect a follow-up attempt. Anyone who calls later promising to recover your money for an upfront fee is running the second half of the same scam.

Where to Report It

Reporting matters even when recovery is a long shot, because it feeds the data that shapes public warnings and, occasionally, freezes funds in time.

Report the scam to the FTC at ReportFraud.ftc.gov. For anything internet-enabled, also file with the FBI’s Internet Crime Complaint Center. If the person targeted is 60 or older, the DOJ’s National Elder Fraud Hotline offers case-manager help in several languages at 833-372-8311. When elder financial abuse is the concern, contact Adult Protective Services, and call local law enforcement if anyone is in immediate danger.

The Honest Bottom Line

The grandparent scam used to have one genuine weakness, the chance that the voice wouldn’t quite match. Voice cloning closed that gap. What it didn’t change is the part that was always doing the real work. The rush, the secrecy, and the demand for an irreversible payment in the next ten minutes.

So stop trying to win the moment with your ears and win it with a habit instead. A code word your family rehearsed. A callback to a number you already trust. A flat rule that emergency money waits until someone else confirms the story. None of it is high-tech, and that’s the point. The scam is built to beat your judgment in real time, so the answer is to decide in advance, calmly, what you’ll do when the call comes.

If you’re setting this up for an older parent, the companion piece on protecting an aging parent from scams covers how to have that conversation without making anyone feel managed. For the wider plan, see the best identity protection for seniors and how government impersonation scams work, the institutional-authority cousin that sits right alongside this one in the imposter-fraud family.

Tom Reardon spent over 20 years in product and operations at major identity protection providers. He writes at MyScamGuide.com to give consumers the honest picture the industry’s marketing never did.

Recommended resources: