Deepfakes and Voice-Clone Scams: How to Protect Yourself and Your Family
AI can now copy a voice from a few seconds of audio. Learn how these scams work, the one habit that beats them, and how to check if an image or video is real.
A phone rings. It sounds exactly like your son: "Mom, I'm in trouble, please send money right now." In some cases, it is not your son at all. Criminals use AI tools that can copy a voice from a short clip, such as a video on social media. This is called a voice-clone scam.
Deepfakes in simple words
A deepfake is a photo, audio clip, or video created or changed by AI to make it look like someone said or did something they did not. Some are harmless fun. Others are used for fraud, harassment, and spreading false news.
Scams to watch for
- Family emergency calls. A cloned voice asks for urgent money after a fake accident or arrest.
- Boss impersonation. A fake voice or video call from a manager asks for an urgent transfer.
- Fake celebrity or expert endorsements promoting investments or miracle cures.
- Romance and friendship scams that use AI-generated photos or videos.
- Fake "official" video calls, including so-called digital arrest calls, where criminals pretend to be police or investigators.
6 habits that protect you
Agree a family code word
Choose a private word or question that only your family knows. Ask for it in any emergency call that asks for money.
Hang up and call back
End the call and phone the person on the number you already have, not the one that called you.
Slow down
Scammers want panic. A real emergency can wait two minutes for a check.
Never send money, OTPs, or gift cards on a call
Especially not through payment links or unknown accounts.
Limit public voice and video
Public clips are raw material for cloning. Set social media accounts to private if you can.
Verify the source of viral media
Before sharing a shocking video, check whether trusted news outlets report it.
How to check an image or video
- Look at the context. Who posted it first? Is there a reliable source?
- Do a reverse image search with tools like Google Lens or TinEye to find the original.
- Check for Content Credentials. Some cameras and tools attach a tamper-evident record, called C2PA Content Credentials, showing how an image was made or edited. It is a growing standard, though not yet on all media.
- Look for small oddities, such as strange hands, mismatched shadows, odd text, or lips out of sync. These are clues, not proof, since AI keeps improving.
If you are targeted
Stop and do not pay
Do not send anything.
Tell someone
Speak to family or a trusted friend.
Report it
In India call 1930 or report at cybercrime.gov.in. Also report the account or message to the platform.
What most people miss
- Anyone can be fooled. Voice clones can fool people who know the speaker well.
- You do not need to be famous to be cloned. A short public video can be enough.
