AI Deepfake Scam Defrauds Italian Bank of Nearly €100M

Fideuram – Intesa Sanpaolo Private Banking, one of Italy’s largest wealth-management groups, was defrauded of nearly €100 million in February 2026 after cybercriminals used AI-powered deepfakes to impersonate senior executives. The attackers leveraged WhatsApp and deepfake voice clones to authorize fraudulent transactions, highlighting the rapid evolution of cybercrime in the AI era.

The Scam

According to TechRadar, cybercriminals targeted Fideuram – Intesa Sanpaolo Private Banking in February 2026 using AI-powered deepfake technology. The attackers impersonated Carlo Messina, CEO of the parent company Intesa Sanpaolo, via WhatsApp. They sent a message to Paolo Molesini, then chairman of Fideuram, requesting assistance with an urgent overseas transaction.

To reinforce the deception, the fraudsters used an AI-generated deepfake of a lawyer’s voice during a follow-up call. Molesini, convinced the request was legitimate, authorized a series of transactions totaling nearly €100 million to foreign accounts, primarily in China and Hong Kong.

Recovery Efforts and Aftermath

Law enforcement agencies in China, Portugal, and Italy collaborated to recover approximately $60.14 million of the stolen funds. However, $40.8 million remains missing, with reports indicating the funds were likely converted into cryptocurrencies—a common tactic to evade recovery.

Paolo Molesini resigned from his roles as chairman of Fideuram and Intesa Sanpaolo Private Banking in March 2026. He cited 'personal reasons' for his departure, though the timing has fueled speculation about potential reputational consequences.

While no executives at Fideuram are under investigation, authorities are reportedly pursuing a 'foreign national living outside Europe' for potential involvement in the fraud.

Broader Implications

This incident is part of a growing trend of AI-enhanced cybercrime, particularly in Business Email Compromise (BEC) attacks. Traditionally, BEC attacks relied on email spoofing, but advancements in AI have enabled fraudsters to exploit instant messaging, voice calls, and even video deepfakes to impersonate trusted individuals.

A similar case involved Italian businessman Massimo Moratti, who was tricked into transferring approximately $1.13 million after scammers used an AI-generated deepfake of an Italian minister’s voice. These cases underscore the need for businesses to update their security protocols to address the risks posed by AI-driven fraud.

What this means

Lazyfounder analysis — our interpretation, not reported fact.

This attack highlights the escalating threat of AI-driven fraud in financial services. For founders and operators, it’s a clear signal that traditional security measures—such as relying on voice or messaging authentication—are no longer sufficient. The use of AI to create convincing deepfakes means businesses must adopt multi-factor authentication (MFA) and real-time transaction monitoring, especially for high-value transfers.

The partial recovery of funds also underscores the challenges of tracking and retrieving stolen assets, particularly when cryptocurrencies are involved. This incident may prompt wealth-management firms to rethink their protocols for authorizing transactions, including stricter verification processes for requests involving senior executives.

Finally, the resignation of Paolo Molesini, while officially attributed to 'personal reasons,' suggests that leaders may face indirect consequences even when they are not directly at fault. This could set a precedent for how companies handle reputational risks tied to cybersecurity failures.

Key takeaways

• Fideuram – Intesa Sanpaolo Private Banking lost nearly €100 million in an AI-powered deepfake scam in February 2026.
• The attackers used WhatsApp and deepfake voice clones to impersonate senior executives and authorize fraudulent transactions.
• $60.14 million was recovered, but $40.8 million remains missing, likely converted into cryptocurrencies.
• Paolo Molesini resigned as chairman in March 2026, citing 'personal reasons.'
• No Fideuram executives are under investigation, but authorities are pursuing a foreign national for potential involvement.

FAQ

What is an AI-powered deepfake?

An AI-powered deepfake is a synthetic media creation that uses artificial intelligence to generate realistic audio, video, or images of a person. In this case, fraudsters used AI to create a convincing voice clone of a lawyer to deceive the victim.

How did the attackers trick Paolo Molesini?

The attackers sent a WhatsApp message to Molesini, impersonating Carlo Messina, the CEO of Intesa Sanpaolo. They followed up with a phone call using an AI-generated deepfake of a lawyer’s voice to convince him the transaction request was legitimate.

How much of the stolen money was recovered?

Approximately $60.14 million of the nearly €100 million stolen was recovered through collaboration with law enforcement agencies. However, $40.8 million remains missing, likely converted into cryptocurrencies.

Why did Paolo Molesini resign?

Paolo Molesini resigned in March 2026, citing 'personal reasons.' While no direct link to the scam has been established, the timing of his resignation has raised questions about potential reputational repercussions.

What are Business Email Compromise (BEC) attacks?

Business Email Compromise (BEC) attacks are a type of cybercrime where fraudsters impersonate executives or trusted partners to trick employees into authorizing fraudulent transactions. AI has enabled these attacks to evolve beyond email to include voice, video, and instant messaging.

Related on Lazyfounder

• More business news