Crypto ATM Fraud in the US Reaches $333 Million in 2025 Amid AI Deepfake Scams

Crypto ATM fraud in the United States surged to $333 million in 2025, with the FBI receiving a 33% increase in complaints as scam networks became more sophisticated, utilizing advanced AI deepfake technology. According to Cointelegraph, cybersecurity firm CertiK highlighted in its latest report that criminal organizations are exploiting the speed and pseudonymity of crypto ATMs, or kiosks, to extract funds from victims at an accelerating pace. The FBI recorded over 12,000 complaints between January and November 2025, marking a significant rise from the previous year. The United States hosts 78% of the world's 45,000 cryptocurrency machines, as noted by CertiK. These machines' ability to convert cash to crypto in under five minutes with minimal identity verification makes them an attractive target for scammers. The report also pointed out an 'attribution gap' due to the blockchain only recording the operator-to-destination transfer, not the victim’s identity, complicating forensic tracing without court orders for operator records. Approximately 86% of losses involve victims over 60, as older adults are particularly vulnerable due to liquid savings, lower crypto literacy, and social isolation. However, younger victims are increasingly targeted in romance or investment scams, commonly referred to as 'pig butchering,' which is one of five primary tactics used by scammers. The other methods include government impersonation, tech support fraud, 'grandparent scams,' and fake fraud recovery offers. Unlike phishing or wallet-draining attacks, which involve compromising private keys or tricking users into signing malicious smart contract requests, ATM-based fraud relies entirely on social engineering to induce the victim to perform a voluntary physical action at a kiosk, stated CertiK. AI-enabled social engineering scams were reported to be 4.5 times more profitable than traditional methods in 2025, according to CertiK. The integration of real-time deepfake synthetic media into scam and fraud operations represents a significant near-term escalation. AI-driven personalization tools enable scammers to scrape social media data and construct hyper-targeted scripts that mimic the specific language, appearance, and communication patterns of the victim's trusted contacts. The profile of crypto ATM scammers has shifted from independent actors to structured transnational criminal organizations operating with corporate-level divisions of labor. Wyoming Senator Cynthia Lummis expressed hope in September that upcoming crypto market structure legislation will help address ATM fraud by punishing bad actors without stifling innovation. In February 2025, U.S. Senator Dick Durbin introduced the Crypto ATM Fraud Prevention Act, aiming to implement safeguards for crypto kiosk users.