Deepfake Voice / AI Voice Cloning Scam
A fraud that uses AI-generated voice audio to impersonate a known person — such as a family member, executive, or public official — in a phone call to extract money or data.
Also known as: AI voice scam, voice cloning fraud, synthetic voice fraud, deepfake audio scam
Last reviewed: 10 June 2026
Advances in text-to-speech and voice-cloning technology now allow fraudsters to create convincing audio imitations of real individuals using only a few seconds of recorded speech. These deepfake voices are deployed in phone calls to family members ('grandparent scam' variants), employees receiving urgent instructions from a fake CEO, or high-pressure scenarios where the victim believes they recognise the voice.
Because the voice sounds familiar and the call may include accurate personal details harvested through OSINT, victims have a strong emotional response that overrides scepticism. Unlike a text scam, a voice call that sounds like a loved one in distress is extraordinarily persuasive.
Establish a family codeword to verify the identity of callers in an apparent emergency. Organisations should mandate out-of-band confirmation for any financial instruction received by phone, regardless of how convincing the caller sounds.
Examples
- An elderly person receives a call from what sounds exactly like their grandson saying he has been arrested and needs bail money wired urgently.
- A finance employee hears the CFO's voice on a call authorising an urgent transfer; it is an AI-generated clone assembled from public videos.