--- title: "Her Daughter's Voice Was Cloned. The Kidnapping Was Fake." published: 2026-07-17T00:03:15.942Z updated: 2026-10-11T06:12:57.396Z author: "Trust Onion Editorial Team" canonical: https://trustonion.io/blog/ai-voice-cloning-fake-kidnapping-scam-savi-security --- # Her Daughter's Voice Was Cloned. The Kidnapping Was Fake. **Cite as:** Her Daughter's Voice Was Cloned. The Kidnapping Was Fake., Trust Onion, 2026-10-11. https://trustonion.io/blog/ai-voice-cloning-fake-kidnapping-scam-savi-security A mother in Arizona picked up the phone and heard her daughter's voice, terrified, claiming she'd been kidnapped. The phone number matched her daughter's. The voice matched her daughter's. Neither was real. Scammers had cloned the daughter's voice using AI and spoofed her number, staging a fake crisis to extort money. The mother believed it, for a few devastating minutes, before the ruse unraveled. What happened to her is no longer unusual. The FTC received more than 36,000 reports of impersonator scams involving family members in 2024, with losses topping $1.1 billion. Voice cloning now requires as little as 20 to 30 seconds of audio, pulled from a voicemail, a TikTok, a YouTube video, or a recorded call. Once scammers have a sample, they can generate speech that sounds like your child, your parent, your spouse, and make that voice say anything. Then they spoof the real phone number. The caller ID says it's your daughter. The voice sounds like your daughter. The message is that your daughter is in danger. ## The Playbook The call comes in, often late at night or during a moment of distraction. The fake voice delivers a few seconds of panic. A second person, the supposed captor, gets on the line with demands. They tell you not to hang up, not to call anyone else, not to contact police, keeping you on the line so you can't verify anything. Pressure plus isolation plus a voice you trust. Detection tools are one response. Patrick and Ryan Coughlin watched this happen to their mother, then built Savi Security, a consumer app with real-time call monitoring and text filtering, and raised $7 million in seed funding to bring it to market. Real-time AI detection during a phone call is genuinely hard. Phone audio is compressed, calls are short, and scammers keep improving. Human accuracy at spotting AI-generated audio hovers around 55 to 60 percent, barely better than a coin flip, and you're not making that judgment in a lab with headphones. You're making it in three seconds, while your heart is pounding. Asking "does this voice sound real?" is not a reliable test. --- There's a simpler question. Instead of trying to detect a cloned voice, you ask something the fake voice cannot answer: "What are the words?" That's how Trust Onion works. Families share three private codewords, known only to the people in their circle, calculated locally on each person's phone. A cloned voice, no matter how convincing it sounds, doesn't know your family's three words. A scammer calling from a spoofed number doesn't know them either, because they were never sent to a server, never stored anywhere that could be breached. The words rotate every 60 seconds, so even if someone somehow learned them, they'd expire almost immediately. Note: if a callback to a known number goes unanswered or sounds wrong, that alone does not confirm the original caller was legitimate. If you believe someone may genuinely be in danger, contact authorities. ## Before the Call Comes Every protection tool has one thing in common: it has to exist before the crisis. Codewords need to be shared during a calm moment, not while someone is screaming on the other end of the line. The conversation is simple: "Hey, if you ever get a weird call from someone claiming to be me, here's what to ask." The Coughlin family's mother had no tool like this when it happened. She had a phone, a caller ID that said her daughter, and a voice that sounded like her daughter. There was nothing to ask, no way to verify, just fear and a demand for money. "What are the words?" would have ended it in seconds. ## A controlled voice-cloning safety demonstration [Try our tool to see just how easy voice cloning is](/clone-my-voice). This controlled safety demonstration exists only to show how easy voice cloning is and to make the risk tangible. It is not a real incident or proof of anyone's identity. Your submitted sample is handled transiently to generate the fixed demonstration audio. Trust Onion does not identify you or the person from the sample. Trust Onion does not store the submitted recording, clone, or generated audio. ## Key takeaways - Voice cloning requires only 20 to 30 seconds of audio from public sources like TikTok or voicemails to convincingly impersonate a family member. - Human accuracy at identifying AI-generated audio is only 55 to 60 percent, so you cannot rely on your instincts to spot a fake voice under pressure. - A family codeword system offers a simple defense that bypasses voice detection entirely, because a cloned voice has no way to know words that exist only on your family's phones. ## Frequently asked questions ### What is a virtual kidnapping scam? A virtual kidnapping scam is a fraud where criminals call and claim to have a family member held hostage. Modern versions use AI-cloned voices to make the victim sound real and scared, pressuring you to pay a ransom before you can verify anything. ### How do scammers clone someone's voice for a phone scam? Scammers collect 20 to 30 seconds of audio from public sources like social media videos, voicemails, or recorded calls. AI tools then generate new speech in that person's voice, allowing criminals to make the cloned voice say anything they want. ### Can you tell if a voice on the phone is AI-generated? Probably not reliably. Studies show humans correctly identify AI-cloned audio only 55 to 60 percent of the time under controlled conditions. In a real, high-stress call you have seconds to decide, making accurate detection even less likely. ### What is a family codeword and how does it stop AI voice scams? A family codeword is a secret phrase shared only among trusted family members. If a caller cannot provide the codeword, you know something is wrong. It bypasses voice detection entirely because a cloned voice has no way to know the secret phrase. ### How common are AI voice cloning scams in 2024 and 2025? The FTC received over 36,000 reports of impersonator scams involving family members in 2024, with total losses exceeding 1.1 billion dollars. Experts expect these numbers to rise as voice cloning tools become cheaper and easier to use. ## More from Trust Onion Protect your family with Trust Onion's three rotating codewords. AI can clone a voice, but it cannot know words that exist only on your family's phones. Trust Onion is free, works offline, and takes about two minutes to set up at trustonion.io. **Safety note:** If you cannot reach the person or organization when you call back, do not assume the original caller was really them. For someone in your trusted group, ask them to give all three current words first and check privately in your app; never reveal your words to an unverified caller. For a bank or other organization outside your group, verify through the official app or a number from a card or statement. Use your best judgment, and alert the authorities if you believe someone is in danger.