CaraComp
CaraComp
Forensic-Grade AI Face Recognition for:
Get Started7-day refund guarantee**
Podcast

Deepfake: Scammers Clone Loved Ones' Faces and Voices

Deepfake: Scammers Clone Loved Ones' Faces and Voices

Quick answer

How do scammers clone a loved one's face and voice with deepfakes?

Scammers use artificial intelligence to copy a relative's face and voice from a short clip or photo posted online, then contact the victim and ask for money or banking details. Because people rarely spot these fakes, the safest step is to hang up and call back on a number you already know.

Deepfake: Scammers Clone Loved Ones' Faces and Voices

0:00-0:00

This episode is based on our article:

Read the full article →

Deepfake: Scammers Clone Loved Ones' Faces and Voices

Full Episode Transcript


Sri Lanka's police just warned the public that scammers can build a fake version of your mother's voice from a short clip on social media.


A few seconds of audio

A few seconds of audio. One photo.

That's often all it takes to make a call that sounds exactly like someone you love, asking for money.

If you've ever posted a birthday video, a voice note, or a family photo, this story is about you.

It's unsettling, and that reaction makes sense.

Here's what police in Sri Lanka are saying.

Criminals use A.I. to copy the faces and voices of relatives, friends, and people victims trust.

Then they contact the victim and push them to send money or hand over banking details.

The Daily Mirror in Sri Lanka reported the warning, but this isn't only a Sri Lankan problem.


What happens when seeing and hearing someone stops

So what happens when seeing and hearing someone stops being proof that it's them?

Start with how fast this is growing.

According to fraud-tracking figures cited in the research, fake media made up a tiny sliver of fraud attempts in twenty twenty-two.

About one in a thousand.

By twenty twenty-five, that was roughly six and a half in every hundred.

That's a jump of dozens of times over, in just a few years.

Now look at one place where it shows up.

The security firm Pindrop measured deepfake fraud attempts at customer call centers.


Trusted by Investigators Worldwide
Run Forensic-Grade Comparisons in Seconds
Detailed facial comparison reports. Results in seconds.
Get Started
7-day refund guarantee**

They went from about one a month to about seven

They went from about one a month to about seven every day.

Imagine being the person on the phone at a bank, hearing a voice that sounds exactly right.

Every instinct says it's real.

And those instincts are the target.

Normally, trust works like this: I know you, so I recognize you.

Deepfakes attack that habit directly.

They fake the very thing your brain uses to say, "Yes, that's them."

Can people spot the fakes?


The research brief points to a test where only

The research brief points to a test where only about one in a thousand people correctly identified A.I. content.

So if you've ever thought, "I'd know a fake if I saw one," the numbers say otherwise.

That's not a personal failing. The fakes are simply that good.

For investigators, it changes the job.

A video or a voice recording can't carry a conclusion by itself anymore.

It has to be checked by some independent means first.

INTERPOL has also pointed to the growing use of A.I. in fraud, with criminals making convincing fake videos and voice recordings to impersonate real people.

For the rest of us, it means a familiar face on a screen is now a claim, not a fact.


There's a useful distinction here

There's a useful distinction here.

Recognizing a face is a human feeling. It can be fooled.

Proving identity is different.

It means measuring and comparing the geometry of a face across multiple images, in a systematic way.

The research describes this as using distance measurements between facial features, applied consistently, instead of a gut feeling.

The idea is that a method beats a hunch.

There's a fair counterpoint.

Detection tools are improving fast.


The Bottom Line

And extra verification steps can slow down legitimate work.

But the brief notes that even trained eyes sometimes can't tell anymore.

So the default has to shift toward checking first.

The flip is this. The danger isn't that fakes are getting better. It's that "looks and sounds right" was never a real test in the first place. We just never had to notice, until now.

Scammers can now copy a face and a voice from a few seconds of material. Most people can't tell the difference by looking or listening. So the safe move is to verify through a second channel, like hanging up and calling back on a number you already know.

Whether you work a case or just answer your phone, this changes what "real" means, and a simple habit of checking gives you back control.

The full story's in the description if you want the deep dive.

Ready for forensic-grade facial comparison?

Full forensic reports with detailed similarity scoring. Results in seconds.

Run My First Search