AI-Enabled Fraud Costs Major Italian Bank $108 Million

Fraudsters stole about $108 million (€95 million) from one of Italy’s largest banks in a scheme aided by artificial intelligence, highlighting the risks financial institutions face as deception becomes harder to detect.

The theft centred on a single phone call. The available details do not identify the bank or explain precisely how AI was used, leaving key questions about the impersonation and authorisation process unanswered.

AI can make phone-based fraud more convincing by imitating a trusted person’s voice. A familiar-sounding caller may pressure staff to bypass routine checks, but a voice alone is not reliable proof of identity.

Banks can reduce that risk by independently verifying sensitive instructions. Calling back through an established contact number, requiring additional approval for large transfers and keeping identity checks separate from the original call can help prevent a convincing impersonation from becoming a costly payment.

Staff training and transaction monitoring are also important safeguards. The central challenge is not simply detecting an artificial voice, but ensuring that no urgent request or apparently authoritative caller can override the controls protecting customers’ money.