RoundTalk

Scammers used AI voices to impersonate CEOs in $129M fraud

English · C2 worksheet Based on the RoundTalk episode of

Free for classroom use. Students listen to the first 75 seconds and read the beginning of the debate, then work on the tasks. The answer key is on the last page. The panelists are fictional characters with AI voices.

Episode page
1 · Listen (first 75 seconds)

2 · Read

Host

Welcome to RoundTalk. Three men allegedly stole $129 million using AI voices to impersonate CEOs. They're caught, but thousands of companies fell for it. How did they pull it off, and why are our defenses so brittle? Marcus, corporate security consultant; Sarah, AI ethics researcher; David, bank compliance officer.

Host

Marcus, you've spent eighteen years watching finance teams under pressure. What made Soall, Hazan, and Sheetrit's calls so effective?

Marcus Webb

The tools are beside the point. What made those calls work was the second voice—the fake British lawyer at a big firm, blessing the wire. Authority stacked on authority. That's where it breaks down.

Host

Sarah, Marcus says the toolkit's almost irrelevant—it's the social engineering that scales. Did Soall and his partners build their own voice cloner, or just licence something off-the-shelf?

Sarah Chen

Bought, almost certainly. Nobody in that indictment is described as an engineer. And that's the problem nobody wants to talk about—somebody built that cloner, somebody sold it, and they're not in the dock in Manhattan. Soall rented a voice. July 2024 to February 2026, undetected.

Host

David, you've trained hundreds of staff on social engineering. When a voice sounds that convincing—the CEO, then the lawyer—do your teams even have a fighting chance?

David Okonkwo

A fighting chance, yes—but not by ear. Last year a clerk of mine got a "CFO" call, perfect voice. She hung up and rang the desk extension. That's it. Callback on a known number, every time.

Marcus Webb

No, wait—callback assumes she thinks something's wrong.

Host

Right, Sarah.

Sarah Chen

That's exactly why it can't depend on suspicion. Make the callback unconditional—every wire, every time, no judgement call. David's clerk shouldn't need a hunch. Soall's whole business model was manufacturing urgency so nobody thinks anything's wrong.

David Okonkwo

Wrong premise, Marcus. Callback isn't a suspicion tool, it's a plumbing tool. My clerk didn't smell a rat—she followed the step. Make it a rule and you don't need instinct.

Key vocabulary

  • impersonate — pretend to be someone else to deceive them
  • fintech — financial technology company or sector
  • undetected — not discovered or noticed by anyone
  • callback — return call to verify identity by ringing back
  • deepfake — synthetic media made with AI to imitate someone

Comprehension questions

  1. Why does Marcus Webb argue that the AI voice technology itself was not the main factor in the fraud's success?
  2. What does Sarah Chen identify as a significant gap in how the perpetrators were prosecuted?
  3. How did David Okonkwo's clerk respond when she received the fraudulent CFO call?
  4. Why does Sarah Chen argue that callback verification cannot rely on employees suspecting something is wrong?
  5. What distinction does David Okonkwo make between how he describes callback verification and how Marcus Webb interprets it?

True or false?

  1. Sarah Chen describes the voice cloning technology as custom-built by one of the three indicted men. T / F
  2. Marcus Webb claims that the perpetrators' success depended partly on impersonating both a CEO and a lawyer from an established firm. T / F
  3. Sarah Chen argues that unconditional callback verification should be standard procedure rather than depending on employees' suspicion. T / F

Fill the gaps

Callbackimpersonateundetected
  1. Three men allegedly stole $129 million using AI voices to   CEOs.
  2. July 2024 to February 2026,  .
  3.   on a known number, every time.

Discuss in pairs

  1. Do you think employees can reasonably be trained to verify every single wire transfer by callback, or would this procedure become impractical in fast-moving financial institutions? What might be the downside?
  2. Marcus Webb and Sarah Chen disagree about whether callback verification assumes suspicion. Where do you stand—is a fixed rule system better than training people to spot warning signs, or should companies use both approaches together?
  3. The person who sold the voice-cloning tool faces no charges. Should developers and sellers of technology bear legal responsibility when criminals misuse their products, or is that an unfair burden?

The whole debate, free every day Students can hear the full 6-minute episode with a synced transcript in the RoundTalk app.

Get it on Google Play Download on the App Store

Answer key

Comprehension
  1. Because the perpetrators stacked authority by adding a fake British lawyer from a major firm to endorse the wire transfers, which proved more persuasive than the voice alone.
  2. The person who actually built the voice cloner and sold it to Soall was never charged, even though they enabled the fraud.
  3. She hung up and telephoned the CEO's desk extension directly to verify the call came from a legitimate number.
  4. Because Soall's business model deliberately manufactured urgency to prevent people from questioning anything, so they cannot be expected to detect fraud through intuition alone.
  5. David frames it as a procedural rule that requires no judgment or suspicion, whereas Marcus assumes it depends on employees first sensing something is amiss.
True or false
  1. False
  2. True
  3. True
Gaps
  1. impersonate
  2. undetected
  3. Callback

roundtalk · https://www.getroundtalk.com/listen/english/c2/scammers-used-ai-voices-to-impersonate-ceos-in-129m-fraud/worksheet/