Technology expert Juan Guevara demonstrates how artificial intelligence can clone a person’s voice by analyzing characteristics such as tone, rhythm, pronunciation and vocal inflection. Once trained, an AI system can generate new speech that sounds similar to the original speaker, even when that person never said those words.

While voice cloning has legitimate applications in entertainment, accessibility, translation and personal digital assistants, the technology can also be misused for impersonation, fraud and extortion. Criminals may obtain audio samples from social media videos, podcasts or other publicly available content and use them to create convincing messages designed to exploit a victim’s sense of urgency or fear.

Guevara recommends that families establish a private verification word or question that can be used whenever a suspicious call or message appears to come from a loved one. As AI-generated audio becomes increasingly realistic, simple identity-verification habits can become an important part of protecting personal and financial information.