A familiar voice can now be weaponized.
Mexico City, May 2026. Criminal groups are increasingly using artificial intelligence to clone voices and commit fraud, creating phone calls that sound like relatives, friends or trusted contacts in moments of supposed emergency. The method is effective because it attacks the victim through emotion before reason can intervene.
The scheme usually begins with a short audio sample taken from social media, videos, voice notes or public recordings. AI tools can then generate a highly realistic imitation, allowing criminals to simulate distress, fear or urgency. Once the victim believes the voice is authentic, the pressure begins.
The most common script is simple and brutal. The caller claims to be kidnapped, detained, injured or trapped in a medical emergency, then asks for immediate money. The objective is not technical sophistication alone, but psychological acceleration: create panic, isolate the victim and force a fast transfer before verification occurs.
Prevention now requires a new family protocol. Relatives should agree on a private verification phrase, avoid reacting to urgent payment requests by phone, and always call back through a known number before sending money. Banks, messaging platforms and families must treat voice as a vulnerable credential, not as proof of identity.
The deeper risk is cultural. For decades, recognizing someone’s voice was treated as intimate evidence of truth. Artificial intelligence has broken that assumption. In the new fraud economy, trust is no longer only stolen through passwords; it can be copied from the sound of someone you love.
Contra la propaganda, memoria. / Against propaganda, memory.