AI News Page

FIELD GUIDE · CLONED VOICES, REAL SCAMS

CAN YOU TRUST A VOICE ANYMORE?

A familiar voice used to mean a familiar person. Cheap cloning turned that old instinct into a security hole.

UPDATED 2026-09-22

TL;DR

THE STORY SO FAR

01

22,000 COMPLAINTS, $893M GONE

The FBI's 2025 Internet Crime report logged 22,000+ complaints with an AI nexus and nearly $893 million in reported losses across AI-enabled scam categories, voice cloning included. That's the whole synthetic-identity swamp, not voice alone.

02

AI VOICES ARE ALREADY USED IN TARGETED IMPERSONATION

The FBI warned of campaigns using AI-generated voice messages to impersonate senior officials — building trust first, then going for account access or fraud. Elite social engineering no longer needs a live con artist.

03

NO SINGLE TECH FIX, SAYS THE FTC

The FTC has dug into watermarking, liveness checks, detection, authentication, upstream safeguards — and warns every fix has holes and the bad guys adapt. Detection can't carry this alone.

04

U.S. TELECOM RULES ALREADY TREAT AI-GENERATED VOICES AS ARTIFICIAL

In 2024 the FCC ruled AI-generated voices count as artificial or prerecorded calls under robocall law. The law already treats synthetic voice as its own risk category.

KEEP READING

MORE QUESTIONS

SHOW YOUR WORK

SOURCES

Cloning has legit uses. Identity checks can't pretend it doesn't exist.

  1. 22,000 COMPLAINTS, $893M GONE2025 IC3 Annual ReportFederal Bureau of Investigation
  2. AI VOICES ARE ALREADY USED IN TARGETED IMPERSONATIONSenior U.S. Officials Impersonated in Malicious Messaging CampaignFederal Bureau of Investigation
  3. NO SINGLE TECH FIX, SAYS THE FTCApproaches to Address AI-enabled Voice CloningFederal Trade Commission
  4. RECOGNITION ALONE IS NO LONGER AUTHENTICATIONScammers Use Fake Emergencies To Steal Your MoneyFederal Trade Commission
  5. U.S. TELECOM RULES ALREADY TREAT AI-GENERATED VOICES AS ARTIFICIALFCC Makes AI-Generated Voices in Robocalls IllegalFederal Communications Commission