AI Can Imitate Family Voices! If You Receive a 'Friend Asking for Money' Call, Do This One Thing First: Full Guide to Identifying AI Voice Scams

"Mom, I'm in trouble and need money urgently—please transfer it right away!" What if the voice on the other end of the phone sounds exactly like your child, parent, or partner? With the advancement of generative AI technology, fraudsters can now forge not only images but also realistic human voices. Combined with fabricated emergencies like car accidents, arrests, or investment losses, these scams can easily trigger panic and cloud judgment. Taiwan's National Police Agency warns that scammers may use 'AI deepfake videos' or 'voice cloning' to demand money transfers, verification codes, or personal information. Therefore, relying solely on 'whether the voice sounds like the person' is no longer a reliable way to verify identity.

If you receive an unexpected call from a loved one asking for money or help, the most important step isn't to scrutinize whether the voice sounds slightly off—it's to immediately end the call and contact the person directly using a known, trusted contact method from your own records. The U.S. Federal Trade Commission (FTC) explicitly warns that even if the voice sounds identical to a family member, you should never trust it alone. Instead, always re-contact the person using a phone number you already know is theirs.

What Is an AI Voice Scam? How Can It Mimic Family Voices?

Previously common 'Guess who I am' scams involved fraudsters impersonating relatives, extracting information through conversation, and then requesting money. Now, AI voice cloning technology makes impersonation far more convincing. The FTC notes that scammers may obtain a short audio clip of a relative—such as from publicly shared online videos—and use voice replication tools to generate a voice nearly identical to the real person. This means that even a familiar voice on the phone cannot guarantee the speaker is actually who they claim to be.

Taiwanese authorities have also issued warnings about such tactics. The Kaohsiung City Criminal Investigation Bureau used AI to alter voices and videos to simulate a scam scenario where a 'partner asks for money after investment losses,' reminding the public that 'seeing is no longer believing, and hearing is no longer proof.' If someone claiming to be a relative asks for money via LINE, always verify by calling their original number.

One Critical Step When Receiving a 'Friend Asking for Money' Call: Hang Up and Verify via Another Channel

If the caller suddenly cries, 'I'm in trouble,' 'I need money urgently,' or 'Don't tell anyone,' the most dangerous response is to immediately comply with the request just because the voice sounds authentic. The key is to resist emotional manipulation and time pressure, end the call immediately, and contact the person directly using a phone number from your own contact list. The FTC stresses that even if the voice sounds exactly like your family member, you should verify using a number you know is correct. If you can't reach the person immediately, check with other family members or friends.

Taiwan police echo this principle. The National Police Agency advises that when receiving video or voice calls from someone claiming to be a relative or official, always verify through another trusted method—such as a phone call or in-person meeting—rather than trusting the caller's media alone.

5 Warning Signs of AI Voice Scams

Warning Sign 1: A Sudden 'Extremely Urgent' Incident

'I hit someone with my car,' 'I got arrested,' 'I'm in the hospital,' 'I lost money on an investment and need a loan now.' A core tactic of these scams is creating time pressure. The FTC notes that scammers often claim the situation is urgent and only the recipient can help, aiming to exploit anxiety and fear to prompt immediate payment without verification. Therefore, if a 'relative' calls unexpectedly with a combination of emergency + urgent money need + immediate payment demand, pause and verify first.

Warning Sign 2: Insisting You Don't Tell Other Family Members

'Don't tell Mom and Dad,' 'This is embarrassing—don't tell anyone,' 'Just handle it yourself.' This is another red flag. The FTC points out that scammers often demand secrecy because involving others increases the chance the scam will be exposed. The more the caller pressures you to keep it secret, the more you should immediately contact other family members to cross-check.

Warning Sign 3: Suddenly Changing Phone Numbers or LINE Accounts Before Asking for Money

'This is my new number,' 'My old LINE isn't working—add my new account.' If a relative suddenly switches contact methods and then immediately asks for money, be alert. The Kaohsiung police warn that common tactics include claiming a number change and asking to be added on LINE or Facebook before requesting money—now enhanced with AI deepfakes to make video calls appear more authentic. Always verify via the original number before sending money.

The FBI also advises that if someone you know contacts you from a new number or platform, re-verify their identity using a previously confirmed contact method.

Warning Sign 4: Failing to Answer 'Family Passwords' or Personal Questions

If you're unsure whether the caller is real, add an extra verification step. Police recommend families pre-agree on a 'secret phrase' to use in suspicious situations. You can also ask personal questions only the real person would know. The FBI suggests setting a secret word or phrase in advance for identity verification. Avoid using easily discoverable info like birthdays, phone numbers, or pet names, which may be public on social media.

Warning Sign 5: Tone or Word Choice Is Unusual

AI-generated voices may still show subtle signs, such as call delays or speech patterns differing from the person's usual style. The FBI advises watching for delays, intonation, and phrasing. However, it also warns that AI-generated content is often 'extremely difficult to detect,' with cloned voices sounding nearly identical to real ones. Therefore, voice anomalies should only be secondary clues—never assume 'normal voice = real person.'

In short, the most reliable defense against AI voice scams isn't trying to out-hear the AI, but verifying identity through an independent channel.

How Families Can Prevent AI Voice Scams: Pre-Agree on a 'Family Password'

Instead of waiting for a scam to happen, families should proactively set up defenses. Agree on a 'family password' and establish a simple rule: whenever a sudden request for money, transfer, or emergency action arises, no matter how authentic the voice sounds, always re-verify first. Both the Kaohsiung police and the FBI recommend using pre-agreed secret phrases for verification.

For elderly relatives unfamiliar with AI, skip technical explanations. Just say: 'If I ever call asking for money, even if I sound exactly like myself, hang up first and call my usual number to confirm.' This rule is far more practical than teaching them to detect AI voices.

Not Just Money: Never Share OTPs, Bank Passwords

AI impersonation isn't only about direct money transfers. The National Police Agency warns scammers may use deepfake videos or cloned voices to request transfers, verification codes, or personal data. Never let realistic-looking media lower your guard. Never share personal info, passwords, or one-time passwords (OTPs). The FBI also warns against sharing two-factor authentication codes with anyone contacting via email, SMS, or messaging apps, as this could lead to account takeover.

Therefore, if a conversation involves money transfers, bank accounts, passwords, OTPs, or other sensitive data, raise your guard immediately.

What If You Already Sent Money? Call 165 Immediately

If you hang up, contact the real person through your usual method, and realize you've been scammed—and already sent money—don't give up. The Criminal Investigation Bureau urges victims to immediately call the anti-fraud hotline 165, report the case, and provide the recipient account and amount. Police can initiate bank alerts to freeze funds. The sooner you report, the higher the chance of stopping the transfer. After calling 165, file a formal report at the nearest police station as soon as possible.

The biggest change AI brings is that the old adage 'seeing is believing' and 'hearing is believing' can no longer be trusted. Next time a familiar voice calls with claims of an emergency, secrecy, account changes, or urgent transfers, don't try to prove whether it's AI. Just do this: end the current contact and reach the real person through your own trusted method.

Sources: This article's information on AI voice cloning, characteristics of family emergency scams, and verification methods is primarily based on public anti-fraud guidance from the U.S. Federal Trade Commission (FTC). AI voice detection, independent verification, and family password recommendations are based on public information from the U.S. Federal Bureau of Investigation (FBI).