Abstract:
A high-end financial fraud case using artificial intelligence technology has recently shocked the European banking industry. Fideuram, a well-known Italian private bank, transferred a total of 95 million euros to criminal gangs due to a well-planned AI voice fraud. Although most of the money was recovered through the cooperation of law enforcement agencies and banks in many countries, approximately 36 million euros are still missing.
Fideuram is the private banking arm of Intesa Sanpaolo, one of Italy's largest banking groups. It mainly provides wealth management services to high-net-worth clients. According to disclosed information, this fraud first began in February 2026.
At the time, Fideuram’s then-chairman Paolo Molesini received a WhatsApp message that appeared to be from Intesa Sanpaolo CEO Carlo Messina. The content of the message asked him to help handle an urgent overseas transaction.
Shortly after receiving the message, Molesini received another call. The caller claimed to be a senior partner of a well-known law firm and further confirmed the authenticity of the relevant transaction arrangements.
However, an investigation revealed that the person on the other end of the call was not actually an attorney.
Informed sources revealed that the criminal gang used artificial intelligence voice cloning technology to copy the lawyer's voice, making it sound almost indistinguishable from his own. By forging high-level instructions and endorsements from legal professionals, fraudsters successfully built a highly credible chain of deception.
To further enhance the authenticity, the fraud gang also sent 11 documents to Molesini, including confidentiality agreements and so-called authorization documents with Messina's signature.
After continuous multiple information verifications, Molesini finally believed that the transaction was real and asked the financial department to arrange the transfer of funds.
Subsequently, Fideuram remitted a total of 95 million euros to multiple overseas accounts, mainly flowing to relevant accounts in China and Hong Kong.
However, the bank's internal team later discovered anomalies in the relevant transactions and quickly launched investigation procedures. At the same time, Fideuram immediately contacted the bank that received the funds and relevant law enforcement agencies and requested to freeze the flow of funds.
With the assistance of relevant institutions in Italy, Portugal and China, the bank finally successfully recovered approximately 53 million euros in funds.
However, the remaining funds gradually disappeared after complicated cross-border account transfers. Investigators believe the remainder of the funds were eventually converted into cryptocurrency, further complicating tracing efforts.
Currently, Italian prosecutors have launched an investigation into a foreign citizen living outside Europe on suspicion of computer fraud. However, so far, no members of Fideuram’s management have been named as targets of the investigation.
It is worth noting that Molesini resigned as chairman of Fideuram in March this year. At that time, he only cited personal reasons as an explanation, and the bank did not provide further explanation on the background of his resignation.
Analysts believe that this case shows that artificial intelligence technology is increasingly being used for high-end social engineering attacks. In the past, fraudsters mainly relied on forged emails and text messages to deceive. Now, with the help of voice cloning, deep forgery and generative AI tools, they have been able to impersonate the true identities of corporate executives, lawyers and other key people.
In fact, similar cases have begun to appear in Italy. Just last year, a fraud gang used AI technology to imitate the voice of an Italian government minister and successfully tricked the well-known entrepreneur Massimo Moratti into transferring nearly 1 million euros to an overseas account. Although the funds were eventually recovered, the incident has exposed the real threat of artificial intelligence fraud.
Industry experts warn that as voice generation technology continues to advance, traditional financial verification mechanisms that rely solely on phone calls to confirm identity are losing their reliability. In the future, financial institutions and large enterprises may need to establish more stringent multi-verification systems to prevent new risks brought by deep forgery technology.
This 95 million euro fraud case has also become one of the largest known AI voice fraud cases. It once again highlights that while artificial intelligence creates huge commercial value, it is also providing unprecedented new tools for cybercrime.
Comments