Warm-up
1. You get a voice message from your CEO asking for an urgent, confidential payment. Which of your instincts would you trust, and which would you doubt?
2. Retrieval: from the Louvre heist and the virtual-influencer texts, which words or collocations could describe a modern fraud? Jot down three before you read.
3. What one check would you personally make before wiring money after an unusual request - and could a fraudster fake it?
Think it through silently or jot notes; you will use your answers in the speaking task.
The Call That Looked Real
recycled new The highlights show the language you will practise.
For decades, the standard advice to anyone handling company money was simple: if in doubt, see the person face to face. In January 2024, that advice was shown to be dangerously flawed. A finance employee at the Hong Kong office of the engineering group Arup received a message, allegedly from the chief financial officer in London, about a confidential transaction. He suspected phishing, but his doubts faded when he joined a video call and saw the CFO and several colleagues. Every one of them is believed to have been a deepfake, built from publicly available footage. Little did he know that he was the only real person on the call. He went on to make fifteen transfers, worth about HK$200 million (roughly US$25.6 million), to five local bank accounts. Only when he contacted head office did the truth emerge.
Not every attempt succeeds. In July 2024, a Ferrari executive received WhatsApp messages from an unfamiliar number, apparently from the CEO, Benedetto Vigna, hinting at a major acquisition. A call followed, and the imitation of Vigna's southern Italian accent was reportedly almost perfect. What set alarm bells ringing was a request for a confidential currency-hedging transaction. Rather than comply, the executive asked a question no impostor could answer: which book had Vigna recommended a few days earlier? No sooner had the question been asked than the call ended. It was a single personal question, not a software tool, that thwarted the scam.
Italy has had its own wake-up call. In February 2025, Defence Minister Guido Crosetto warned that fraudsters had cloned his voice to phone prominent businessmen, asking for money to free kidnapped journalists. Most targets saw through the ruse, but the former Inter Milan owner Massimo Moratti wired almost a million euros. Police later tracked down the funds and froze them.
What links these cases is psychology rather than technological brilliance. Fraudsters impersonate authority figures, tailor each message to the victim, create a sense of urgency and insist on secrecy; our brains do the rest, because expectation shapes perception and we hear what we expect to hear. A familiar face or voice will gain credibility in seconds. In Hong Kong, the criminals pulled off their scheme without compromising a single internal system, Arup said: they simply bypassed the people. Experts therefore argue that protocols, not instincts, are the best way to mitigate the risk.
Per decenni il consiglio standard per chi gestiva il denaro aziendale era semplice: nel dubbio, incontra la persona di persona. Nel gennaio 2024 quel consiglio si è rivelato pericolosamente inadeguato. Un dipendente dell'area finanza della sede di Hong Kong del gruppo di ingegneria Arup ricevette un messaggio, presumibilmente dal direttore finanziario di Londra, relativo a una transazione riservata. Sospettò un tentativo di phishing, ma i suoi dubbi svanirono quando si collegò a una videochiamata e vide il CFO e diversi colleghi. Si ritiene che ognuno di loro fosse un deepfake, costruito a partire da filmati pubblicamente disponibili. Non sapeva di essere l'unica persona reale nella chiamata. Eseguì quindici bonifici, per circa 200 milioni di dollari di Hong Kong (circa 25,6 milioni di dollari USA), verso cinque conti locali. Solo quando contattò la sede centrale la verità venne a galla.
Non tutti i tentativi riescono. Nel luglio 2024 un dirigente di Ferrari ricevette messaggi WhatsApp da un numero sconosciuto, apparentemente dell'amministratore delegato Benedetto Vigna, che accennavano a una grande acquisizione. Seguì una telefonata, e l'imitazione dell'accento meridionale di Vigna era, secondo quanto riferito, quasi perfetta. Ciò che fece suonare il campanello d'allarme fu la richiesta di un'operazione riservata di copertura sul cambio. Invece di assecondare, il dirigente pose una domanda a cui nessun impostore avrebbe saputo rispondere: quale libro gli aveva consigliato Vigna pochi giorni prima? Appena fu posta la domanda, la chiamata si interruppe. Fu una semplice domanda personale, non un software, a sventare la truffa.
Anche l'Italia ha avuto il suo campanello d'allarme. Nel febbraio 2025 il ministro della Difesa Guido Crosetto avvertì che dei truffatori avevano clonato la sua voce per telefonare a noti imprenditori, chiedendo denaro per liberare dei giornalisti rapiti. La maggior parte dei bersagli smascherò il raggiro, ma l'ex proprietario dell'Inter Massimo Moratti versò quasi un milione di euro. La polizia in seguito rintracciò i fondi e li congelò.
Ciò che accomuna questi casi è la psicologia più che la genialità tecnologica. I truffatori impersonano figure di autorità, adattano ogni messaggio alla vittima, creano un senso di urgenza e insistono sulla riservatezza; al resto pensa il nostro cervello, perché l'aspettativa plasma la percezione e sentiamo ciò che ci aspettiamo di sentire. Una voce o un volto familiari guadagnano credibilità in pochi secondi. A Hong Kong i truffatori portarono a termine il piano senza compromettere alcun sistema interno, ha dichiarato Arup: aggirarono semplicemente le persone. Gli esperti sostengono quindi che i protocolli, non l'istinto, siano il modo migliore per attenuare il rischio.
Recycled language
Items you have already met in earlier lessons. Notice them working inside the new text.
| Item | Meaning | In this lesson | First met in |
|---|---|---|---|
| pull off (a scheme) | succeed in doing something difficult or risky | The criminals pulled off their scheme without hacking a system. | Louvre Heist |
| bypass | go round a barrier or control | They bypassed the people, not the software. | Louvre Heist |
| protocol | official procedure | Protocols, not instincts, are the best defence. | Louvre Heist |
| flawed | containing faults | The old advice was dangerously flawed. | Louvre Heist |
| track down | find after searching | Police tracked down the funds and froze them. | Louvre Heist |
| gain credibility | become trusted | A familiar voice gains credibility in seconds. | Virtual influencers |
| tailor the message | adapt it to one person or group | Fraudsters tailor each message to the victim. | Virtual influencers |
| mitigate (the risk) | reduce how serious something is | Protocols mitigate the risk. | ShotSpotter / Space law |
| allegedly | claimed but not proved | A message allegedly from the CFO. | 20 World Secrets |
| is believed to have + past participle | passive reporting of earlier events | Every one of them is believed to have been a deepfake. | Conspiracy Theories HW |
| It was X that... / What X was... | cleft sentences for focus | It was a single question that thwarted the scam. | Contract Emphasis |
| Only when/after... + auxiliary + subject | restrictive inversion | Only when he contacted head office did the truth emerge. | Contract Emphasis |
| Under no circumstances shall... / Should + subject... | negative and conditional inversion | See the protocol below. | Contract Emphasis |
| expectation shapes perception | top-down processing | We hear what we expect to hear. | Hypnosis, Brain & Perception |
| I have my doubts about... / Where is the evidence of...? | language of doubt | Used in Exercises 6, 7 and the speaking task. | Conspiracy Theories HW |
| narrative past perfect | an event earlier than the story's 'now' | ...which book had Vigna recommended? | Crime & Punishment 2 |
New language
| New item | Meaning | Example |
|---|---|---|
| deepfake | video or audio made by AI to imitate a real person | The call was full of deepfakes. |
| impostor / impersonate | a person who pretends to be someone else / to pretend to be | No impostor could answer that. |
| wire (funds) / wire transfer | send money electronically | Moratti wired almost a million euros. |
| thwart | prevent a plan from succeeding | A personal question thwarted the scam. |
| red flag | a warning sign | Urgency plus secrecy is a red flag. |
| set alarm bells ringing | make someone suddenly suspicious | The request set alarm bells ringing. |
| see through (a ruse) | realise that something is a trick | Most targets saw through the ruse. |
| a sense of urgency | pressure to act quickly | Fraudsters create a sense of urgency. |
| call back on a known number | verify by phoning a number you already have | Call back on a known number. |
| dual authorisation | two approvers for one payment | Large transfers need dual authorisation. |
| Little did + subject + verb | inversion for dramatic narrative ('he had no idea') | Little did he know he was alone. |
| No sooner had + S + pp + than... | one event immediately after another | No sooner had the question been asked than the call ended. |
Practical protocol
- Under no circumstances shall a payment be released on the strength of a video call alone.
- Should a request arrive from an unfamiliar number, call back on a known number before doing anything.
- Treat urgency plus secrecy as a red flag: ask a question only the real person could answer.
- Large transfers need dual authorisation. It is the second approver who catches what the first one missed.
- Questioning a senior colleague's request is professionalism, not insubordination.
Myth vs reality
Reality: The Arup employee suspected phishing and still joined the call - the video removed his doubts. Careful people are targeted precisely because they look for confirmation.
Reality: Real-time faces and cloned voices can now imitate people, accents included. Seeing and hearing are no longer proof of identity.
Reality: Arup said no internal system was compromised. The weak point was human trust, not the network.
Exercises
1. Image memory game
Matches: 0/8 - Attempts: 0
2. Complete the sentence
3. True or false?
4. Which case?
5. Structure recognition
6. Everyday workplace scenarios
7. Integration dialogue
8. Written production
Words: 0 (target 120-150)
Output tasks
Brief your CFO. In one minute, tell the group which of the three cases worries you most and why. Use at least: allegedly, protocol, mitigate, and one cleft sentence (It was... that...).
Say it aloud: 'Under no circumstances shall...' - finish the sentence for a rule at your own company. Then add 'Should...' with a second rule.
Find ONE recent deepfake or voice-clone case (any country, 2025-26). Write five sentences about it: two with passive reporting (is said / believed / alleged to...), one with 'Little did...' or 'No sooner had...than', and one with a cleft. Bring the link.
Sources
- Finance worker pays out $25 million after video call with deepfake CFO (CNN, Feb 2024, via AI Incident Database)
- Ferrari exec foils deepfake attempt by asking a question only the CEO could answer (Fortune / Bloomberg, 27 Jul 2024)
- Italian police freeze cash from AI-voice scam that targeted business leaders (Reuters, 12 Feb 2025, via AI Incident Database)
- Arup deepfake loss: no internal systems compromised (CFO Dive)
Texts above are original summaries written for this lesson from the cited reports; check the links for the full coverage.