How to Spot AI-Generated Scams and Deepfakes
AI has made fraud cheaper and more convincing. Messages are grammatically perfect, voices can be cloned from a few seconds of audio and fake videos look increasingly real. The good news is that the underlying tricks are old, and simple verification habits still work.
Common AI-powered scams
- Voice-cloning calls: a relative or boss “urgently” asks for money.
- Deepfake video calls: an executive appears to approve a transfer.
- Polished phishing: emails and texts with no obvious typos.
- Fake investment ads: celebrity “endorsements” generated by AI.
- Romance and support scams: chatbots that keep a conversation going for weeks.
Red flags that never go out of style
Urgency, secrecy, unusual payment methods such as gift cards or crypto, pressure to bypass normal procedure, and requests that arrive from a new number or email address.
Spotting a deepfake
Look for unnatural blinking, lip movements that do not match speech, odd lighting or blurry edges around the face and hair, and audio that sounds flat. None of these is proof, and quality keeps improving, so rely on verification rather than detection.
Verification habits that work
- Hang up and call back on a known number.
- Agree a family code word for emergencies.
- Require two approvals for payments at work.
- Check news and official sites before trusting a viral video.
- Never act on a request just because it sounds like someone you know.
If you were targeted
Stop contact, do not send further money, report to your bank immediately, change passwords and report the scam to your local fraud agency and the platform used.
