DOJ indicts trio for $129M CEO deepfake fraud using voice and image cloning

Court filings detail Soall scripting lawyer impersonations while Hazan and Sheetrit handled laundering through layered accounts in the US, Europe, UAE, and China before crypto conversion. Specific losses include $4.9M from a Missouri hospital and $11.5M from a UK tech firm. The scheme layered AI-generated CEO calls with follow-on attorney spoofs to bypass verification protocols. Hazan’s 2008 bank fraud conviction and claimed $400M laundering history since 1994 form the operational backbone rather than novel AI alone. Indictment evidence rests on transaction records, communication logs, and arrest locations in Spain and the UK rather than disclosed forensic artifacts proving AI model provenance. Official statements attribute the scale to emerging tech; independent review shows continuity with prior business email compromise patterns where social engineering, not model sophistication, drove success. Cross-border movement of funds to crypto mixers mirrors documented mule networks in earlier Israeli-linked fraud rings. Procurement and regulatory filings reveal no new defensive mandates for voice biometric thresholds despite repeated warnings from financial institutions. Victims spanned hospitals and banks lacking standardized callback or multi-channel confirmation. Next phase hinges on extradition outcomes and whether additional crypto exchange records surface further nodes before 2025 asset seizures.