AI Voice Clone and Fake WhatsApp Message Used in €95M Bank Fraud
Scammers impersonated bank CEO using AI to authorize fraudulent transfers from Italian private bank.
1Sophisticated AI-powered fraud
The scam targeted Paolo Molesini, chairman of Fideuram (Intesa Sanpaolo's private banking arm), combining a fake WhatsApp message purportedly from the bank's CEO with an AI-generated voice clone of a senior lawyer. This two-pronged social engineering attack convinced Molesini to authorize €95 million in transfers to foreign accounts. The fraud demonstrates how AI tools are being weaponized against financial institutions' verification processes, exploiting human trust in familiar voices and hierarchical authority.
2Multi-stage deception tactics
The fraud began in February 2026 with a WhatsApp message appearing to come from Intesa Sanpaolo CEO Carlo Messina, requesting assistance with an overseas transaction. The scammers then reinforced this request with a phone call using AI voice cloning to impersonate a senior law firm partner. This combination of digital impersonation and synthetic voice technology created a convincing illusion of legitimacy, bypassing standard verification protocols at the private banking institution that typically serves high-net-worth clients.
3Cross-border money trail
Fideuram transferred €95 million to accounts in China and Hong Kong before detecting anomalies. Through international cooperation with authorities in Italy, Portugal, and China, €53 million was recovered. The remaining €42 million was laundered through complex overseas account networks and converted to cryptocurrency, making full recovery unlikely. This case highlights the challenges of tracing fraudulent transactions across jurisdictions with varying regulatory frameworks and cooperation protocols.
4Executive consequences
Molesini resigned in March 2026 without public explanation, though the fraud investigation likely prompted his departure. As chairman of a private bank serving affluent clients, his position involved overseeing personalized financial management services where trust and security are paramount. The incident raises questions about executive accountability in fraud cases involving sophisticated social engineering, where human judgment intersects with technological deception.
5Institutional vulnerabilities
Fideuram's private banking model, which emphasizes personalized service and direct executive involvement in client transactions, may have created vulnerabilities exploited by the scammers. The fraud targeted the human element in financial authorization processes rather than technical systems. This incident exposes how traditional verification methods in private banking may be inadequate against AI-enhanced social engineering attacks that mimic trusted voices and hierarchical chains of command.
6Law enforcement response
Milan prosecutors are investigating the international fraud ring, with one suspect outside the EU already under scrutiny for computer fraud. The case demonstrates how cybercriminals are leveraging AI tools to execute complex, multi-jurisdictional financial crimes. Authorities face challenges in attribution and prosecution due to the transnational nature of the crime and the use of cryptocurrency for final money laundering stages.
7Broader industry implications
The Fideuram case represents one of the largest known AI-assisted financial frauds, signaling a new era of cybercrime threats to banking institutions. Financial regulators worldwide are likely to reassess authentication protocols for high-value transactions, particularly in private banking where personalized service models may conflict with robust security requirements. The incident underscores the need for multi-factor authentication systems that can detect synthetic media.
8Emerging security paradigm
This €95 million fraud establishes AI voice cloning as a credible threat to financial institutions, requiring new defensive measures beyond traditional cybersecurity. Banks must now authenticate not just the content of communications but the very nature of the communicators themselves. The financial industry faces an arms race against increasingly sophisticated AI tools that can mimic human behavior patterns and exploit institutional trust structures.
Original article: TechSpot — AI Voice Clone and Fake WhatsApp Message Used in €95M Bank Fraud