The FBI says AI voice-clone kidnapping calls helped drain $893 million from Americans in a single year

Federal investigators are warning that criminals are using artificial intelligence to copy the voices of family members and stage fake emergencies over the phone. In one increasingly common version of the scheme, a caller plays what sounds like a relative in distress and demands money to resolve a supposed kidnapping or arrest, exploiting panic to push victims into paying before they can verify anything.

How a cloned voice fuels the panic

Voice-cloning tools can now reproduce a recognizable imitation of a person from only a short sample of recorded speech, and social media clips or voicemail greetings can supply that sample. Scammers use the synthetic audio to make a plea for help sound authentic, then layer on urgency by claiming the relative has been taken or injured and that time is running out.

The Federal Bureau of Investigation has warned that these AI voice-cloning scams mimic loved ones in distress, and it has tied the surge to hundreds of millions of dollars in losses, according to a summary of the bureau’s warning. The emotional design of the calls is the point: a frightened person is far less likely to pause and check whether the emergency is real. Callers often shout, cry or hand the phone to a supposed captor, tactics meant to keep the target reacting rather than thinking.

The scale of the losses

The financial toll attached to AI-related fraud is substantial. The bureau reported that Americans lost more than $893 million in a single year to scams involving artificial intelligence, a category that includes voice-clone impersonation alongside other AI-assisted deception. The figure reflects reported losses, which experts note typically understate the true total because many victims never come forward, whether from embarrassment or a belief that reporting will not help.

Losses of that size point to a shift in how quickly and convincingly fraud can be manufactured. Techniques that once required a skilled impersonator can now be produced with widely available software, lowering the barrier for criminals and widening the pool of potential targets. The same tools that let a hobbyist recreate a voice for entertainment can be turned toward deception, and the gap between a rough imitation and a convincing one has narrowed sharply. Because a usable sample can be scraped from a public video or a brief phone conversation, almost anyone with an online presence can supply the raw material, and the volume of ready audio circulating on the internet gives criminals an ample supply to draw from.

Why older adults are frequent targets

Impersonation scams often lean on family bonds, and callers frequently pose as a child or grandchild in trouble. That framing can be especially effective against older adults, who may be less familiar with how convincingly voices can now be faked and more inclined to act immediately when a grandchild seems to be in danger.

Criminals also research their targets, pulling names and relationships from social media to make a call feel personalized. When the voice sounds right and the details line up, the deception becomes harder to detect in the moment the decision to pay is being made. A scammer who can name a grandchild, reference a recent trip or mimic a familiar speech pattern removes many of the doubts that would otherwise give a target pause.

The verification steps investigators recommend

Authorities emphasize slowing down as the single most protective response. Because the calls rely on urgency, hanging up and independently contacting the relative through a known phone number can expose the ruse before any money changes hands. If the family member answers safely, the emergency was fabricated.

Investigators also suggest agreeing on a private code word within a family, a shared phrase that a real relative could supply and an impersonator could not. Treating any unexpected demand for immediate payment, particularly by wire transfer, gift cards or cryptocurrency, as a warning sign is another line of defense, since those payment methods are difficult to reverse once sent. A legitimate emergency almost never hinges on paying a stranger in gift cards within minutes, so that specific demand is itself a strong signal that a call is fraudulent.

How the technology outpaced old defenses

For years, the sound of a familiar voice served as a kind of built-in proof of identity, and much of everyday trust rested on that assumption. Synthetic audio erodes that safeguard, because a voice can no longer be taken as conclusive evidence that the speaker is who they claim to be. The result is that families and institutions alike are being pushed to adopt verification habits that do not depend on how a caller sounds.

Some banks and companies have begun advising customers to confirm unusual requests through a second channel, and consumer-protection officials have folded voice cloning into broader warnings about impersonation fraud. The underlying lesson is that a convincing voice is now something that can be manufactured, and defenses have to account for that rather than assume the phone line carries an honest signal.

Reporting and what comes next

Officials encourage victims and near-victims to report incidents so law enforcement can track patterns and pursue the networks behind them. Reporting also helps agencies measure a threat that is evolving quickly as the underlying technology improves and spreads, and it feeds the data that produces figures such as the $893 million total.

The broader concern is that convincing synthetic audio is no longer rare or expensive to produce, which means the tactic is likely to persist. Public awareness remains one of the few defenses that scales, because a caller counting on shock and speed loses power the moment the person on the other end knows to stop and confirm. Warnings that circulate before a family is targeted tend to be the most valuable, since a person who has already heard that voices can be faked is far more likely to hesitate when a frantic call arrives. That simple pause, investigators say, is often all it takes to defeat a scheme built entirely on manufactured urgency.

This article was produced with the assistance of AI and reviewed by Morning Overview editors prior to publication.

More from Morning Overview