AI-enabled scams cause millions in losses across US and Africa
Scammers are using generative AI to clone voices, create deepfakes and spoof phone numbers to commit fraud affecting consumers across the United States and South Africa. The FBI reported nearly $893 million in AI-enabled scam losses in 2025, with the technology making traditional scams significantly harder to recognize.
What happened
Scammers are using generative AI to clone voices, create deepfake images and videos, and spoof phone numbers to commit fraud. According to the FBI's 2025 Internet Crime Report, Americans reported nearly $893 million in AI-enabled scam losses, with a Cobb County couple losing roughly $800,000 in a cryptocurrency scam and a South Florida immigration attorney discovering criminals had cloned his voice to target immigrants.
AI is being used to make traditional scams more convincing by cloning voices from social media, creating fake photographs, researching family members, and making emergency requests appear authentic. The scams include grandparent scams using voice cloning, romance schemes, fake investment opportunities, and distress scams.
South Africa has also reported increasingly sophisticated banking scams that may involve international criminal syndicates operating from Cambodia and are facilitated by AI and data harvesting techniques.
Reported impact
Florida reported 350 AI-related scam complaints totaling nearly $39.9 million in losses, while Georgia reported 71 complaints with more than $10.4 million in losses. The FBI notes these figures likely undercount the true scope because many victims never report being scammed.
Romance and confidence scams resulted in more than $19 million in reported losses nationwide, and AI-assisted distress scams including voice-cloning schemes resulted in more than $5 million in losses.
Index analysis
AI-related scam complaints remain a fraction of all reported internet crimes but the financial impact continues to grow. The technology is making familiar scams harder to recognize by increasing their believability and sophistication.
Classification
Sources and evidence
This record was researched and written by the Index, based on the following original reporting.