
- AI-powered deepfakes convinced executives to authorize massive overseas fund transfers
- Incident highlights evolving business email compromise tactics in the AI era
Cybercriminals have tricked a major Italian bank into wiring more than $100 million abroad by targeting executives with AI-powered deepfakes. Some of the money has since been recovered, but a significant portion remains unaccounted for.
The target was Fideuram – Intesa Sanpaolo Private Banking, a very large Italian private-banking and wealth-management group owned by Intesa Sanpaolo. It cannot be considered a “normal” retail bank - its core business is managing investments and wealth for affluent and high-net-worth clients.
According to the latest figures, the bank has around $513.25 billion in assets under management (AUM), more than 7,100 private bankers, and more than $9.85 billion in net profit.
Deepfakes and phone calls
Citing two sources familiar with the matter, Reuters reported that the attack took place in February 2026, while the chairman was Paolo Molesini.
Molesini reportedly received a WhatsApp message that seems to have come from Carlo Messina, Intesa Sanpaolo Chief Executive Officer (CEO), who asked him to help with an urgent overseas transaction, to which the victim agreed. He was later contacted by someone claiming to be senior partner at a prominent law firm, to confirm the instruction. This person used AI tools to create a deepfake of the lawyer’s voice, tricking Molesini into thinking the request was authentic.
The chairman then reached out to the bank’s finance department, arranging a series of transactions to foreign account, the majority of which were in mainland China and Hong Kong.
The scam was spotted soon afterwards, and Fideuram moved to block the transaction. With the help of law enforcement in China, Portugal, and Italy, around $60.14 million was recovered. Around $40.8 million remains missing, apparently converted into cryptocurrencies.
Ruining a person's career
Paolo Molesini announced he was stepping down from roles of president for both Fideuram itself and Intesa Sanpaolo Private Banking a month later, in March 2026, citing “personal reasons”.
According to , neither Molesini, or any other Fideuram executives are under investigation, but local authorities are apparently investigating a “foreign national living outside Europe” for potential fraud.
In the pre-AI era, this type of fraud was known as Business Email Compromise (BEC). Crooks would break in (or otherwise spoof) the email address of a company’s CEO (or other executive), map out employees working in the finance department, and reach out with instructions for wire transfers. These instructions would explain that the transfer needs to be completed in secrecy, to avoid unnecessary publicity or competitors.
Nowadays, with the advent of artificial intelligence and the ease with which crooks can create deepfake images, voice clones, and even videos, BEC attacks have spread out to instant messaging platforms, voice calls, and even video calls.
Last year, scammers used artificial intelligence to create a deepfake of the voice of an Italian minister, tricking local businessman Massimo Moratti to transfer around $1.13 million to an overseas account. The money was later successfully recovered, it was said.

