Small Town Sysadmin
Free CJIS checklist

AI guidesSafetyFor everyone7 min read

AI-powered scams: cloned voices, deepfake video and fake bosses

How criminals use AI to impersonate family members, executives and officials, the warning signs, and the simple office habits that stop these scams cold.

Last checked:

The scams themselves aren’t new: the panicked call from a relative, the urgent wire request from “the boss,” the official who needs you to click a link. What AI changes is how convincing they are. A cloned voice can sound exactly like your grandson or your bank president, and the scam email no longer has the spelling mistakes that used to give it away.

The good news: the defenses are simple, and they work no matter how real the voice sounds.

What the FBI and FTC are warning about

In a December 2024 public service announcement (I-120324-PSA), the FBI said criminals are using generative AI to:

  • Write messages and build fake websites that read naturally, without the grammar mistakes people used to spot.
  • Create fake photos and documents, including profile pictures and identification.
  • Clone voices to impersonate a family member in a crisis, or to try to get into bank accounts.
  • Fake video calls that appear to show a company executive, a police officer or another authority figure.

The FTC warned as early as March 2023 that a scammer needs only a short audio clip of someone’s voice, the kind that’s easy to find in videos posted online, to clone it for a fake family-emergency call.

The FBI has also warned, in May and December 2025, about an ongoing campaign of texts and AI-generated voice messages impersonating senior US government officials, aimed at officials and the people they know. The goal is usually to get the target to click a link or move to another messaging app, then steal their accounts.

The warning signs

The voice or face can be perfect, so look at what’s being asked instead:

  • Urgency and secrecy. “Don’t tell anyone.” “I need this in the next hour.” “I’m in a meeting, just handle it.”
  • A payment you don’t normally make, or a change to how a regular payment is made (new bank account, new address).
  • Unusual payment methods: gift cards, cryptocurrency, wire transfers, or payment apps.
  • A request for codes or passwords, including the six-digit code that was just texted to you.
  • A switch of channels. Moving from email to text, or from a phone call to a messaging app you don’t normally use.
  • Odd details in images or video, like distorted hands, strange shadows, or a face that doesn’t quite move right. Don’t rely on spotting these; the fakes keep improving.

Habits that stop these scams

  1. Hang up and call back on a number you already know. Not the number in the message or the caller ID. Use the number in your contacts, on the company website, or on the back of your card. This one habit defeats almost every impersonation scam.
  2. Agree on a code word. Families can pick a secret word or phrase to confirm identity in an emergency. Offices can do the same for urgent payment requests.
  3. Never pay with gift cards, crypto or wire because someone called or texted. Legitimate businesses and government agencies don’t demand payment that way.
  4. Verify every payment change by phone. If a vendor emails new bank details, call them at the number you already had on file before changing anything.
  5. Require two people for money movement. A second person approving wires and new payees stops a single convincing call from becoming a loss.
  6. Never share sign-in codes. No real employee, bank or agency will ask for the code that was just sent to you.
  7. Share less online. Limit who can see your social media posts, videos and voice recordings, and think about what your office website reveals about who approves payments.

If you’ve already sent money, act fast. Call your bank immediately to try to stop or recall the payment, then report it. Speed matters more than embarrassment.

For banks and credit unions

Treasury’s Financial Crimes Enforcement Network (FinCEN) issued alert FIN-2024-Alert004 in November 2024 on deepfake fraud targeting financial institutions. It describes criminals using AI-altered or AI-generated identity documents to open accounts, and deepfake audio or video in phishing and impersonation. Red flags it lists include photos that don’t match the customer’s other information, a customer using a third-party webcam plugin or claiming technical problems during live verification, a customer who declines multifactor authentication, and rapid activity in newly opened accounts. FinCEN asks institutions to use the key term FIN-2024-DEEPFAKEFRAUD in related suspicious activity reports. Confirm current procedures with your BSA officer.

Where to report

For more on everyday scams, see how to spot a scam email and what to do if you think you’ve been hacked.

Sources

Last checked October 3, 2026. AI products change quickly, so this guide is re-checked against official sources every month.

Back to the AI guides