Deepfake Basics

Deepfake CEO Fraud: How Businesses Get Tricked

November 20, 2026•By Alex Morgan
Sunlit office desk with a laptop open on a video call, showing a professional in a suit, while a person in the foreground sits with a notebook, suggesting an unexpected request from a senior colleague

What is deepfake CEO fraud? Scammers clone a boss's voice or face to push urgent payments. Here's how businesses get tricked, and how to stop it.

A message arrives from the finance director. A voice on the phone sounds like the chief executive. A video call shows familiar colleagues asking for a confidential payment today. In a growing number of reported cases, the person making that request is not who they appear to be. This is deepfake CEO fraud: a scam that uses a cloned voice, a fake video, or both, to impersonate a senior leader and push an employee to move money. This guide explains how the scams work, what has been reported in real incidents, and the checks that stop them. For the family version of the same tactic, see voice cloning scams.

What is deepfake CEO fraud?

Deepfake CEO fraud is a form of executive impersonation. Criminals copy a real leader's voice, face, or both, then use that likeness to ask an employee — usually in finance, legal or operations — to make an urgent transfer, change a supplier's bank details, or share access. It is closely related to business email compromise. The difference is the extra layer of proof: a voice or face that seems to belong to the boss, on a live call.

The FBI has warned that criminals use generative AI to make fraud more believable, and that they generate videos for real-time chats with alleged company executives (FBI IC3, December 2024). In an earlier advisory, it described business email compromise schemes that already used virtual meetings, including a still picture of a CEO, or "deep fake" audio, while the caller claimed the camera was not working (FBI IC3, February 2022). Today the face and the voice can both be fabricated, sometimes for several people on the same call. Not every executive scam is a deepfake, but the practical response is the same: do not treat a familiar voice, face or profile photo as proof of identity when the request is unusual, urgent or secret.

How a deepfake CEO scam typically unfolds

Reported cases follow a fairly stable pattern:

  1. Collect public material. Talks, interviews and social posts give attackers enough audio and video to clone a voice or a face. The FBI has noted that short audio clips can be enough to start.
  2. Open a channel that looks official, often an email, a WhatsApp message or a calendar invite that appears to come from a senior leader, sometimes using a publicly available photo as the profile picture.
  3. Create urgency and secrecy. The request is framed as a confidential deal or a payment that cannot wait, so the employee does not call back on a known number.
  4. Add a voice or a video as proof. A phone call, a voice note or a video meeting is used to overcome doubt. On a group call, several fake colleagues can appear at once, which makes the request feel socially confirmed.
  5. Ask for a hard-to-reverse payment, such as a wire transfer, a change to bank details, a currency transaction, or personal documents. Once the money moves, it is quickly sent on to other accounts.

A voice call that sounded like the boss

In 2019, The Wall Street Journal reported that the chief executive of an unnamed UK energy firm was tricked into transferring €220,000 (about $243,000) after a call that sounded like the boss of the company's German parent (Wall Street Journal, August 2019; also reported by Forbes). The information was shared with the Journal by the firm's insurer, Euler Hermes Group. According to that account, the UK chief executive recognised the subtle German accent and the "melody" of his boss's voice, and sent the money to what he was told was a Hungarian supplier.

The fraudster called more than once: first to start the transfer, then to claim it would be reimbursed, and again to seek a further payment. The victim became sceptical when the reimbursement had not arrived and when a later call came from an Austrian number. He did not send a second payment. The first transfer had already moved on, and the company was not publicly named. The lesson was already clear in 2019: a voice that sounds right is not a verification method.

A video call in which every other person was fake

In February 2024, Hong Kong police said a finance worker at a multinational firm had been tricked into paying about HK$200 million (about US$25.6 million) after a video conference with people he believed were the company's chief financial officer and other colleagues. Senior Superintendent Baron Chan Shun-ching told the city's public broadcaster RTHK: "(In the) multi-person video conference, it turns out that everyone [he saw] was fake" (CNN, February 2024).

Police said the worker had first received a message that appeared to come from the company's UK-based chief financial officer, talking about the need for a secret transaction. He initially suspected a phishing email, then set that doubt aside after the video call, because the other people on it had looked and sounded like colleagues he recognised. Hong Kong public broadcaster RTHK, cited by CNN in a later report, said the money was sent across 15 transactions. The fraud was discovered only when the employee later checked with the company's head office.

In May 2024, the London-based design and engineering firm Arup confirmed it was the company involved. A spokesperson told CNN that Arup had notified Hong Kong police in January, and said: "we can confirm that fake voices and images were used." The company also said its financial stability and business operations were not affected and that none of its internal systems were compromised (CNN, May 2024). Singapore's Cyber Security Agency also cited the case (CSA, March 2024). The employee had suspected a phishing email, then been talked out of that caution by several familiar faces and voices. No internal system had to be hacked: a real member of staff authorised the payment.

Attempts that were stopped in time

Not every attempt succeeds, and the unsuccessful ones show what works as a defence. In May 2024, The Guardian reported that fraudsters had targeted WPP, the advertising group, by impersonating its chief executive, Mark Read. According to an email from Read to leadership, obtained by The Guardian, the attackers created a WhatsApp account using a publicly available image of him and used it to set up a Microsoft Teams meeting that appeared to include him and another senior executive. During the meeting, the impostors used a voice clone and YouTube footage, and impersonated Read off-camera in the chat. The target, an "agency leader," was asked to set up a new business in an attempt to solicit money and personal details. "Fortunately the attackers were not successful," Read wrote. A WPP spokesperson confirmed that the incident was prevented. Read also warned: "Just because the account has my photo doesn't mean it's me" (The Guardian, May 2024).

In July 2024, Bloomberg reported a similar attempt at Ferrari. According to people familiar with the episode, a Ferrari executive received unexpected WhatsApp messages that appeared to come from chief executive Benedetto Vigna, from a number that was not Vigna's usual business mobile, talking about a confidential acquisition and a currency-hedge transaction. A follow-up call imitated Vigna's voice, including his southern Italian accent. The executive became suspicious and asked a question only the real person could answer: the title of a book Vigna had recently recommended. The call ended. Bloomberg reported that Ferrari opened an internal investigation; the company declined to comment (Bloomberg, reported in Fortune, July 2024). In both attempts, a profile photo and a familiar voice were not enough; a question or a call-back on a known number was the check that mattered.

Why these scams work, and the signs to treat as a stop

Deepfake CEO fraud exploits ordinary workplace habits: discretion, speed, and routine video meetings. Public talks, which are a legitimate part of a leader's job, also give attackers material to copy. Generative AI, the FBI notes, can also correct for the human errors that might otherwise serve as warning signs of fraud. "Listen for a robotic voice" is not a sufficient policy. Some clones will sound off; some will not. The FBI advises listening to tone and word choice, but treats hanging up and calling back on a known number as the reliable step.

Treat any of these as a reason to pause and verify through a second channel: a secret or unusual transaction; extreme urgency; a new number, account or meeting link; a change to payment details, including a first-time supplier, gift cards or cryptocurrency; a caller who avoids a simple question or ends the call when asked to wait; and odd tone, lag or word choice. The absence of those audio signs does not mean the call is genuine.

How businesses can stop deepfake CEO fraud

The control that matters most is procedural: no payment, bank-detail change or sensitive disclosure based on a single channel, however convincing the voice or video looks.

  • Require a second, known channel for money movement. Confirm any request to transfer funds or change supplier details by calling the requester back on a directory number you already have — not by replying on the same call, chat or email thread.
  • Make "secret transaction" a red flag, not a reason to skip checks. WPP's internal warning, as reported by The Guardian, flagged any mention of a secret acquisition, transaction or payment that no one else knows about.
  • Use a question only the real person could answer, then still call back. A shared phrase or a recent conversation can expose a clone, as in the Ferrari attempt reported by Bloomberg, but some personal details can be scraped or guessed.
  • Train finance and executive assistants together on the same rule: pause, verify, and no penalty for delaying an urgent-sounding request. Do not rely on caller ID or a profile photo; both can be copied or spoofed.

Can you check a suspicious voice note or video?

A detector can be a useful extra signal, but it is not a substitute for a call-back. Be clear about what DeepWall's free checkers can and cannot do today.

The free image checker works now. It accepts a JPG, PNG or WEBP file of up to 4 MB, with no sign-up, and returns two scores from 0 to 100%: one for AI generation and one for face swapping, plus, when available, the AI generator the image most likely came from. A still frame or screenshot from a call can be checked on the deepfake image detection page. Daily limits apply (5 image checks per visitor). Images are analysed in memory and not saved.

You can also upload a short video on the deepfake video detection page (MP4, MOV or WEBM, up to 30 seconds and 50 MB) or an audio file on the deepfake audio detection page (MP3, WAV, M4A, OGG or FLAC, up to 25 MB). Automated analysis of those uploads is still in development, so the video and audio checkers do not yet return a real-or-fake score. Files are held only while they are processed and then deleted. If a payment request is in play, hang up and verify through a trusted channel first. For the wider family of voice scams, see voice cloning scams.

Finding a fake of your leaders before it reaches finance

A one-off upload only helps with a file you already have. Watching for a fake profile, cloned voice or video of a leader online is a different job, and it is the paid product, not the free checker.

DeepWall's deepfake detection software is built for that monitoring. You enroll a person with 3 to 5 photos and a 30-second voice sample. DeepWall then checks newly uploaded videos on YouTube, X, TikTok, Facebook, Instagram, Reddit and news sites with embedded video, matching faces and checking the audio for a cloned voice. When a video crosses your match threshold, you get an alert, typically within minutes of detection. Professional and Executive plans add evidence export and takedown templates; removal decisions are made by each platform. Plans start at $49 per month, with a 30-day free trial and no credit card required. Start your free trial, or try the free deepfake detector first.

What to do if your business has been targeted

  1. Stop contact and do not send any further money, codes or documents.
  2. Call your bank immediately if a payment has gone out, and ask it to attempt a recall.
  3. Preserve the evidence: emails, chat threads, invites, numbers, voice notes, recordings and payment references.
  4. Warn finance, legal and the people whose likeness was used, because the same material may be tried on someone else.
  5. Report it. In the United States, the FBI asks victims to file a complaint at ic3.gov. Elsewhere, report the incident to your local police or national fraud service, and to your insurer if relevant.

Frequently asked questions

What is deepfake CEO fraud?

It is a scam in which criminals impersonate a senior leader with a cloned voice, a fake video, or both, and ask an employee to move money, change payment details or share access.

Is deepfake CEO fraud the same as business email compromise?

It is closely related, and sometimes used as the next step in that scam. The email or chat opens the door; the voice or video is there to make the payment request feel confirmed.

Can a cloned voice or a fake video of a CEO be detected just by listening or watching?

Not reliably. The FBI notes that generative AI can correct for errors that used to give fraud away, and that cloned voices can sound nearly identical to the real person. Hang up and verify the request on a number you already trust.

What should a finance team do if the CEO asks for an urgent transfer on a call?

Pause the payment. Call the executive back on a directory number, or confirm through the normal approvals process, even if the voice and the face seemed right.

Can DeepWall check a CEO voice note or a video-call recording?

You can upload a still image or screenshot to the free image checker and get AI-generation and face-swap scores today. Short video and audio uploads are accepted, but those free checks do not yet return a score. Paid monitoring is the product for impersonation of your executives in newly uploaded public videos.

#deepfake ceo fraud#deepfake ceo scam#business email compromise#voice impersonation