AI scams and deepfakes: when you can't trust your eyes or ears
Cloned voices, fake videos, and AI-written messages are making scams more convincing. The defences are simple — and they don't require spotting the fake.
A few seconds of audio from a social video is enough to clone a voice. A photo is enough to fabricate a video call. AI has removed the old warning signs — bad grammar, robotic voices — so the answer is not to "spot the fake." It is to change how you verify.
The scams to know
- Family-emergency calls: a cloned voice of your child or parent begs urgently for money.
- Celebrity investment videos: deepfaked public figures endorsing crypto platforms that steal deposits.
- Video-call impersonation: fraudsters joining work calls as "the boss" to authorise payments.
- Perfect phishing: AI-written messages in flawless language, personalised from your public posts.
Defences that beat any deepfake
- Verify on a channel you control. Hang up and call the person back on their known number. That single habit defeats voice cloning completely.
- Agree a family safe word for genuine emergencies — something no stranger could know.
- Treat urgency plus money as a stop sign, whatever the voice or face says. Real emergencies survive a five-minute verification.
- Never take investment advice from a video. No legitimate opportunity guarantees returns or arrives via a celebrity clip.
Talk about this with older relatives and teenagers especially — knowing that voices can be faked before a call comes is most of the protection. If money was sent, act fast: bank first, then report via our reporting directory.