Is That Really Your Family on the Video Call? How to Verify

Safety & scams Tutorial6 min read·Updated July 4, 2026
The short answer

Real-time deepfake technology can overlay a convincing fake face and voice onto a live video call. The most reliable defenses are things you set up before any emergency: a family safe word, a personal question only insiders know, and always calling back on a saved number rather than the one shown on your screen.

Your phone rings. It is a video call, and on screen you see what looks like your son or daughter. They are upset — there has been an accident, they need money right now, please don't call anyone else.

This scenario is no longer science fiction. Real-time deepfake tools can put a convincing face and a cloned voice onto a live call. The best defenses are things you set up before any such call ever arrives.

Set up a family safe word today

A safe word is a short, random phrase — something like "pineapple ladder" or "blue carnival" — that every member of your household knows and keeps private. It is not a word anyone would guess, and it is not written anywhere obvious.

Agree on the word in person or during a call you initiate yourself. Tell your family: "If I ever ask for the safe word, just say it. And if I ever call you in an emergency and can't give you the safe word when you ask, don't send money until you've called me back on my real number."

Run a quick visual check at the start of any unexpected call

When you receive an unexpected video call from a family member or colleague asking for something urgent, look carefully at the image before reacting:

Watch the edges of the face — especially around the hair and ears. Deepfakes often show a subtle shimmer or warping at facial boundaries. Ask the person to turn their head slowly to one side. Profile angles are harder for current systems to render cleanly. Look at lighting: if the face appears slightly flatter or more evenly lit than the background, that is a possible sign of an overlay. These checks catch lower-quality systems. Higher-quality deepfakes may pass them — which is why the next steps matter too.

Ask something only the real person would know

A deepfake can replicate someone's face and voice, but it cannot give a scammer knowledge of your private history. Ask a question that only your real family member could answer:

"What did we eat at Dad's birthday dinner last month?" or "What is the name of your cat?" or "What did you call me as a nickname when we were kids?" Keep the question conversational and unexpected — not something that could be researched from their social media profiles.

If they hesitate, get evasive, or say "I don't have time for this" — that is a meaningful signal.

Hang up and call back on a number you saved yourself

Caller ID can be spoofed. A call that appears to come from your daughter's number may not be from her phone at all.

If anything feels off, say "I need to call you back in one minute" and end the call. Then go to your contacts and dial their real number — the one you saved yourself, not the one shown on your incoming call screen. If they answer and the emergency is real, you have lost 60 seconds. If it was a scam, you have protected yourself from a potentially devastating mistake.

A real emergency does not evaporate in the 60 seconds it takes to call back on a verified number.

Notice audio-video sync and background details

Current real-time deepfake systems often struggle with natural mouth movements, especially for sounds like "b," "p," and "m" that require visible lip contact. Watch the mouth carefully as the person speaks — any lag between speech and lip movement is a technical artifact.

Also listen for audio quality changes. If the voice sounds slightly robotic or hollow, that can indicate a cloned voice. Background details can also be revealing — a completely blurred or replaced background on an otherwise normal video call is sometimes used to hide the real environment of whoever is operating the deepfake.

Trust your instincts and make the call-back rule a habit

If a video call feels wrong in any way you cannot quite name, trust that feeling. Emotional manipulation — urgency, fear, embarrassment, love — is exactly what scammers rely on to bypass your better judgment.

The call-back rule is simple: for any unexpected video or voice call asking you to take a financial action, always hang up and call back on a verified number before doing anything. Make this a standing rule for your whole family. It takes one minute and it stops this entire category of scam.

What to Try Next

For a broader understanding of how AI-generated video fakes work and what visual tells to look for in recorded content, how to spot a deepfake video goes deeper on the technical side. And if you have not set up a family safe word yet, the family safe words guide walks through the setup in full detail.

Published July 4, 2026 · Updated July 4, 2026How we test →

Frequently asked questions

Can you really fake someone's face on a live video call?
Yes. Tools that run on consumer hardware can replace a face in real time, though quality varies. Most real-time deepfakes still show visible artifacts when the person moves quickly or lighting changes — but this technology is improving.
What is a family safe word and how does it help?
A safe word is a secret word your family agrees on in advance. If someone calls claiming to be a family member in an emergency, ask for the safe word. A scammer won't know it. Your real family member will.
Does the turn-your-head test actually work?
It can reveal lower-quality deepfakes, which struggle with profile angles and fast motion. Higher-quality systems handle this better, so the test is a starting point — not a guarantee. Combine it with other checks.
What if I get an emergency call from someone who looks and sounds like my child?
Stay calm. Say you need to verify it is them before you can help, ask for the safe word, then hang up and call their actual saved number. A real emergency will still be a real emergency 60 seconds later.
Radim Sekera
Founder & editor

Radim is a software developer who spends his days building with AI and his evenings explaining it to family members who don’t care how it works — only what it can do for them. Every guide is tested by hand before it’s published.