Today · Live desk Week of Oct 16–23, 2026 · every story links its source
DEEPFAKECENTRAL
Report a fake Subscribe

Voice-clone scams

How they work, the red flags, and a safe-word plan for your family
← Back to front page

Last updated Oct 23, 2026 · By the DeepfakeCentral desk · Sources linked throughout

The 30-second version

If a call, voicemail or video from someone you love or work for asks for money or secrets in a hurry, hang up and call them back on a number you already know. Agree on a family safe word today. Never pay by wire, gift card or crypto because a voice told you to.

How a voice-clone scam works

Cheap AI tools can now imitate a person's voice from a short recording. The FTC warns that scammers can pull that audio from videos people post online, and the FBI says criminals use AI-generated audio to impersonate loved ones and public figures to demand money.

The safeguards on those tools are often thin. A March 2025 Consumer Reports assessment of six voice-cloning products found that four lacked meaningful safeguards to stop someone from cloning a voice without the speaker's consent.

The technology is new; the script is old. It is the classic impostor scam with a familiar voice attached: create panic, demand secrecy, and push for a payment that is hard to reverse.

Pattern 1: the fake family emergency

A parent or grandparent gets a call that sounds like a child or grandchild: they've been in a crash, arrested, or kidnapped, and need money now. Often a second person takes the phone, posing as a lawyer, police officer or kidnapper. The FTC guide to fake emergencies lists the tells: pressure to act immediately, a request to keep it secret, and payment by wire transfer, gift cards, cryptocurrency or cash handed to a courier.

Variants this month follow the same playbook with different voices. In Florida, police said scammers used AI to clone a sheriff's captain's voice for fake arrest-warrant calls, and Dutch police arrested suspects in an AI-voiced "bank helpdesk" ring.

Pattern 2: the fake boss, the fake bank

In "CEO fraud", an employee gets an urgent request that appears to come from an executive. AI makes it far more convincing. In 2024, engineering firm Arup confirmed that an employee in Hong Kong sent about US$25 million after a video call in which the "CFO" and colleagues were deepfakes, as CNN reported.

The FBI has also warned that since April 2025 criminals have sent AI-generated voice messages impersonating senior US officials to build trust before stealing account access. And fake videos of politicians promoting investments or "welfare payouts" keep circulating, including the deepfake Modi and Sitharaman ads flagged in India this week.

Red flags

  • Urgency: "right now", "before the hearing", "or they'll hurt me".
  • Secrecy: "don't tell Mom", "keep this between us", "this deal is confidential".
  • Hard-to-reverse payment: wire transfer, gift cards, crypto, a payment app, or a courier collecting cash.
  • A switch of channel or number: a new phone number, a new messaging app, or a link to "move the conversation".
  • Caller ID you can't trust: numbers can be spoofed, so a familiar name on screen proves nothing.
  • Dodged questions: the caller gets vague or ends the call when you ask something only the real person would know.

Your safe-word plan

  1. Pick a family safe word or phrase in person, not by text or social media, and make it something no one could guess from your online profiles. The FBI recommends a secret word or phrase to verify identity.
  2. Agree on the rule: anyone asking for money or help in an emergency gets asked for the word. No word, no money.
  3. Hang up and call back on a number you already have for that person, or call another relative who can check on them.
  4. Slow it down. A real emergency survives a five-minute check. Scams depend on you not making that call.
  5. Brief older relatives, who are frequent targets of these schemes, and practice the script once so it feels natural.

For businesses

The NSA, FBI and CISA guidance on deepfake threats urges organizations to plan for impersonation of executives. Practical controls:

  • Require verification through a separate channel for any payment or credential request, using a known number or internal system, never one supplied in the request.
  • Use dual approval for wires and for changes to vendor bank details.
  • Train finance, HR and help desk staff that a convincing voice or face on a call is not authentication.
  • Agree on code words or challenge questions for high-risk requests from executives.

Reduce your exposure

You can't make yourself impossible to clone, but you can make yourself a harder target. The FBI suggests limiting online content of your image and voice, making social accounts private and limiting followers to people you know.

  • Audit public audio and video, such as old public videos, podcasts and voicemail greetings with your full name.
  • Remove yourself from people-search sites. These listings can expose relatives' names, ages and phone numbers. You can opt out site by site for free, or pay a data removal service such as Incogni or Optery to file requests for you. No service can remove everything, and listings can reappear.
  • Consider identity monitoring if your details have been exposed in breaches. Paid services such as Aura bundle credit, identity and account monitoring. They alert you after something happens; they don't stop a voice-clone call.
  • Use a password manager and two-factor authentication so a scammer who tricks you into revealing one password can't take over everything.

We have no affiliate relationship with any product named here. See our affiliate disclosure for how we handle partnerships.

Frequently asked questions

How much audio does a scammer need to clone a voice?

Not much. The FTC warns that a scammer may need only a short audio clip of a family member's voice, which can come from content posted online. That is why the FBI suggests limiting public audio and video of yourself and your family.

What is a family safe word?

A secret word or phrase that only your family knows, agreed in person in advance. If someone calls claiming to be a relative in trouble, ask for it. The FBI recommends this in its December 2024 warning about AI-enabled fraud.

Is it illegal to use an AI-cloned voice in robocalls?

In February 2024 the FCC ruled that AI-generated voices count as "artificial" voices under the Telephone Consumer Protection Act, so robocalls using them need the called person's prior consent and follow the same rules as other robocalls. Using a cloned voice to steal money is fraud regardless.

I sent money to a voice-clone scammer. What now?

Contact your bank, card issuer, wire service or payment app immediately and ask them to stop or reverse the payment. Then report it at ReportFraud.ftc.gov and ic3.gov, file a police report, and save the phone number, times, messages and transaction details.

Will my bank refund me?

Not always. Payments you authorized yourself, such as wires, crypto and gift cards, are much harder to recover than unauthorized card charges. Act fast and ask the bank to treat it as fraud.