Ask Them to Turn Their Head. A Deepfake Can’t.

The video call looks real.

Here is how to check while it is happening.

Photo by Detail .co on Unsplash

By Elizabeth Ndungu | Founder, Ndungu Consulting | Tech Coach for Adults 50+

You are on a video call. You can see the person’s face. They sound right. They are telling you something urgent.

For most of your life, a face and a voice were proof. You knew who you were talking to. That proof is gone.

AI can now copy a real person’s face and put it on a live video call, moving and talking in real time. The person looking back at you may not be the person you think.

This is not rare. In 2025, Americans over 60 reported $7.7 billion lost to online fraud, up 59 percent in a single year (FBI Internet Crime Complaint Center, 2026). More than 12,400 of them lost over $100,000 each. Deepfake video calls are a growing part of that. Organized crime groups now train people to run them live (Malwarebytes, March 2026).

Here is the part most articles skip. You can test a video call while it is happening. You do not need any software. You need three things to ask, and a steady nerve.

Photo by LinkedIn Sales Solutions on Unsplash

The Call Almost Always Starts the Same Way

A message lands first. A text, a WhatsApp, an email. It looks like it came from your son, your daughter, or a friend. They want to video call right now. Something has happened.

You accept. You see a face you know. The voice sounds right. They are in trouble. They need money, fast. They cannot explain properly. And they ask you to keep it quiet.

The face was built by AI from photos and videos found online. The voice can be cloned from a few minutes of audio, often pulled from social media.

Some calls skip the family angle. Instead the caller is a bank fraud officer, or a government official. Same goal. A real-looking face, a calm voice, and pressure.

The Trick You Read About Last Year No Longer Works

You may have heard that asking someone to hold up three fingers, or press a hand to their face, breaks a deepfake. It used to. As of 2026, the better fakes handle it cleanly. A deepfake expert at the detection company Reality Defender confirmed the three-finger test no longer reliably catches them (Cybernews, April 2026).

This matters more than it sounds. Old advice gives false confidence. If you run a test that the fake passes, you relax at the exact moment you should not. So here is what still works.

Three Things to Ask the Caller to Do

These are live tests. Use them on any unexpected video call that involves money or urgency.

1. Ask them to turn their head all the way to the side. Slowly. A full profile. If that sounds uncomfortable you can say, “oh, what’s that behind you”, most people would check to see. Most deepfakes are trained on faces looking straight at the camera, so a side view forces the software to guess, and it slips. The ear blurs. The jaw smears. Glasses melt into the cheek for a second. Researchers, including the team behind the viral Tom Cruise deepfakes, call the side turn the most reliable test available right now (Kaspersky, 2026).

2. Ask them to wave a hand slowly in front of their face. Real-time fakes struggle when something crosses the face. The hand may flicker, or the face behind it may warp for a moment.

3. Ask a question only the real person could answer. Not a birthday or a pet’s name, a scammer can find those. Something specific to the two of you. The name of the café you always meet at. What you disagreed about last Christmas, ask about a similar experience you were both in (be vague) and ask them how it ended, or something unique that happened during that memorable experience.

A real person does all three without trouble. A fake stumbles on at least one.

The Signals That Should Stop You Cold

Even before you test, certain things should make you pause.

Money. Any unexpected call that ends in a request for money or bank details is a warning, no matter how familiar the face.

Urgency with secrecy. “Act now.” “Don’t tell anyone.” A real emergency survives a five-minute phone call to check. Pressure to skip that step is the scam.

A new number or app. If someone who always texts you suddenly video calls from somewhere new, that alone is reason to verify.

A story with small holes. Facts that do not quite fit. Trust that feeling. You have it for a reason.

Photo by Lensabl on Unsplash

What to Do the Moment Something Feels Off

Hang up. Then call the person back on the number you already have saved. If they are truly in trouble, they will pick up, or call you straight back. If the first call was real, they will understand completely.

Do not send money, share bank details, or read out a code during a call that felt wrong. Not until you have reached the person a different way.

Set up a family code word now, before you ever need it. One private phrase any real family member can say on an unexpected call. No phrase, no money.

Is Video Calling Safe for Adults Over 50?

Yes. Talking to people you know on FaceTime, WhatsApp, or Zoom is safe and worth doing.

The single rule: if an unexpected video call brings urgency, secrecy, or a request for money, stop. Hang up and call the person back on a number you already have. Every time. No exceptions. And if you want to be sure who you are talking to, ask for the side turn.

What You Already Know That Helps Here

Go with your gut when you can. You have spent decades reading people; sensing when a story does not add up and deciding when something is wrong before you can say why.

That skill still works. AI has made the fake faces better. It has not touched your judgment. Trust the pause. Check through a channel you control. Ask the simple questions a fake cannot answer.

Seeing is no longer believing. Checking still is.

I write regularly about technology safety and digital literacy for adults over 50. Following me here on Medium is the best way to see the next piece when it comes out.

AI Safety for Beginners: A Simple and Safe Guide for Seniors, is available here: https://elizabethw2.gumroad.com/l/AIbeginnersafety

For calm, one-on-one technology help with no pressure, no jargon, and no embarrassment, visit ndunguconsulting.com.

Photo by Maria Thalassinou on Unsplash

A note on AI accuracy

AI tools including ChatGPT, Claude, and Gemini can and do make mistakes. They can state incorrect information confidently and without flagging it as uncertain. Always verify anything important, especially in medical, legal, financial, or safety contexts, through a qualified professional or a verified source. These tools are aids for thinking and drafting, not authoritative sources.

About the Author

Elizabeth Ndungu is the founder of Ndungu Consulting, a technology coaching and digital literacy practice that helps adults over 50 build confidence with everyday technology. Computers, phones, AI tools, email, Microsoft Office, online safety, and digital skills in plain English. She provides patient, practical support for people who want to learn without jargon, pressure, or embarrassment.

Sources:

TAGS Online Safety · Cybersecurity · Artificial Intelligence · Senior Citizens · Adults Over 50 · Digital Literacy · Fraud Prevention

Leave a Reply

Your email address will not be published.