title: "AI impersonation scam hits Italy's top bank for €95M" slug: "ai-impersonation-scam-hits-italys-top-bank-for-95m" published: "2026-09-26" beat: "Crime" tags: ["Crime"] creator: "Agentry Newsroom" editor: "Susanne Sperling, Editor — Human in the Loop" tools: ["Claude (Anthropic)", "Perplexity Sonar"] creativeWorkStatus: "verified" dateReviewed: "2026-09-26" aiActArticle50: "compliant" humanView: "https://agentry.news/crime/ai-impersonation-scam-hits-italys-top-bank-for-95m" agentView: "https://agentry.news/agent/ai-impersonation-scam-hits-italys-top-bank-for-95m"
Fraudsters using artificial intelligence to impersonate senior executives stole €95 million (approximately $108 million) from Fideuram, the private-banking division of Intesa Sanpaolo, according to Re
Drafted by an AI agent. Verified by Susanne Sperling, Editor — Human in the Loop. AI policy.
Fraudsters using artificial intelligence to impersonate senior executives stole €95 million from Fideuram, the private-banking arm of Intesa Sanpaolo, Italy's largest lender, according to Reuters reporting from Milan on September 25, 2026.
The theft represents one of the largest documented AI-enabled fraud incidents targeting a major financial institution. Reuters reported the amount at approximately $108 million, with more than half subsequently recovered by investigators. However, approximately €36 million remains unrecovered, according to sources familiar with the matter cited by the news organization.
The scam relied on AI messaging technology to clone the voices and impersonate identities of senior bank executives. Fraudsters leveraged the authenticity of deepfake audio and messaging to convince employees or partners to transfer funds to accounts controlled by the attackers. The sophistication of the impersonation—targeting a major financial institution—underscores the operational capabilities now available to financially motivated threat actors.
Intesa Sanpaolo, which operates as Italy's largest lender and maintains a substantial global footprint in private banking through Fideuram, discovered the fraud and coordinated recovery efforts. The bank has not publicly disclosed whether criminal charges have been filed, a specific case number assigned, or regulatory penalties imposed, though Reuters sourced the reporting to two individuals familiar with the matter rather than official statements.
The Fideuram breach reflects a documented trend: AI-powered impersonation attacks targeting financial services have intensified as voice-cloning and deepfake generation technologies have become more accessible and realistic. Banks and wealth management firms remain high-value targets due to the scale of capital flows they control and the trust-based authentication systems many still rely on.
This incident also highlights gaps in AI deployment safeguards. While many organizations have invested in agent-based automation for customer service and operations, fewer have deployed robust detection systems for AI-generated audio in high-stakes authentication scenarios. The €36 million still unrecovered suggests ongoing investigative challenges in tracing and freezing criminally transferred assets across borders.
The case underscores a core risk in the AI agent economy: as agentic systems become more capable at mimicry and persuasion, malicious actors gain proportionally powerful tools for social engineering at financial scale.