AI Voice Cloning Drives $893M in U.S. Fraud Losses in 2025

AI Voice Cloning Drives $893M in U.S. Fraud Losses in 2025

Artificial intelligence has become a powerful tool for criminals targeting American consumers, with AI-enabled scams costing victims nearly $893 million in 2025 according to the FBI's latest Internet Crime Report. The losses represent a sharp escalation in fraud schemes that use voice cloning, synthetic images, and deepfake technology to impersonate trusted individuals.

Recent cases in Georgia and Florida illustrate the sophistication of these attacks. A Cobb County couple lost approximately $800,000 in a cryptocurrency scam, while a South Florida immigration attorney discovered criminals had cloned his voice and identity to target immigrants seeking legal services, CBS News Atlanta reported.

The FBI received more than one million internet crime complaints in 2025, totaling over $20 billion in reported losses across all categories. AI-related complaints represented a small but rapidly growing subset of that total.

Why it matters

The financial impact of AI fraud extends beyond individual victims to erode trust in digital communications at a systemic level. As generative AI tools become cheaper and more accessible, the barrier to launching convincing impersonation attacks continues to fall—forcing organizations and families to rethink how they verify identity in urgent situations. The technology transforms familiar scams into threats that bypass traditional red flags.

Regional impact shows concentration of losses

Florida recorded 350 AI-related complaints with nearly $39.9 million in reported losses, while Georgia saw 71 complaints totaling more than $10.4 million. Federal investigators acknowledge these figures likely undercount the true scope, as many victims never report being scammed.

Shawn McCroy, communications and outreach coordinator for the Georgia Department of Law, said the so-called grandparent scam has become particularly prevalent. In these schemes, criminals use audio samples harvested from social media to generate convincing phone calls that mimic a grandchild or relative claiming to need emergency funds.

How criminals weaponize AI

Authorities identified several techniques scammers now deploy routinely:

• Cloning voices from publicly available social media videos
• Spoofing legitimate phone numbers to appear as trusted contacts
• Creating synthetic photographs that pass casual inspection
• Mining social media profiles to research family relationships and make emergency requests seem authentic

The FBI reported that AI played a growing role in romance and confidence scams, which resulted in more than $19 million in losses nationwide last year. Voice-cloning schemes and other AI-assisted distress scams accounted for an additional $5 million in reported losses.

Truist Bank told CBS News that fraudsters are using artificial intelligence to create increasingly believable scams built around urgency and impersonation, urging customers to pause before acting on unexpected requests.

Recommended defenses

The Georgia Attorney General's Consumer Protection Division recommends families establish a private code word to verify a loved one's identity during emergency calls, noting that AI-generated voices can sound authentic enough to fool even close relatives.

Additional protective measures include verifying urgent requests by contacting the person directly through a known phone number, never sending money based solely on a phone call or voicemail, and reporting suspected scams to local law enforcement and the FBI's Internet Crime Complaint Center.

As AI tools continue to evolve, investigators emphasize that verification—rather than speed—represents consumers' strongest defense against these increasingly sophisticated attacks.

These details were first reported by CBS News Atlanta.