AI voice cloning technology enables sophisticated deepfake scams

Artificial intelligence is increasingly capable of creating convincing voice clones using as little as three to ten seconds of audio. These deepfakes can be harvested from social media or professional content, making them difficult to distinguish from authentic recordings, especially when audio is compressed via phone calls or messaging apps.

Common fraudulent tactics include virtual kidnapping scams, where criminals impersonate a loved one to demand money. To combat this, security experts suggest using pre-agreed ‘safety words’ with family members. Other indicators of synthetic speech include unnatural intonation, unusual pauses, or inconsistent pronunciation. However, because genuine recordings can also suffer from poor quality or echoes, auditory cues alone may not provide absolute certainty.