Protect Against Deepfake Impersonation Fraud in Accounts Payable- Part I

What You Need to Know About the Growing Threat of AI-Generated Deception
Have you tried Oprah’s pink salt weight-loss plan? Well, don’t. According to Malware Tips, that’s not really Oprah.[1] Meanwhile, Tom Hanks has had to publicly announce that, “There’s a video out there promoting some dental plan with an AI version of me. I have nothing to do with it.”
Deepfake impersonations are sophisticated artificial intelligence-generated audio and video content that can convincingly mimic real people’s voices, facial expressions and mannerisms. And they are not confined to celebrities seeming to promote bogus products. Deep-fake impersonations increasingly include CEOs, CFOs and others, directing an accounts payable staffer to make an electronic payment to a vendor or entity, often “confidentially” and rather urgently.
Using advanced machine learning algorithms, criminals can create incredibly realistic but fake recordings of executives, colleagues or business partners. They need just a few minutes of source material from public videos, social media posts or recorded meetings.
The technology has become so advanced that it is very difficult to distinguish between authentic and fabricated content. What once required professional film-level production capabilities can now be created using readily available software and minimal technical expertise.
How Criminals Target Accounts Payable Teams
Of course, accounts payable personnel handle large sums of money that make for lucrative targets. Fraudsters specifically target accounts payable staff because they:
- Have authority to process payments and wire transfers
- Occasionally or regularly receive urgent payment requests from executives
- Usually do not have direct, frequent contact with senior leadership
- Often work under time pressure, making verification less likely
Common Deep Fake Attack Methods:
Common methods of attack include a variety of approaches.
- Fake Executive Calls: Criminals use voice cloning to impersonate a CEO, CFO or other executive to request an urgent wire transfer for a “confidential deal” or “emergency situation.”
- Deepfake Video Conferences: Sophisticated scams use fake video calls with multiple “participants” that appear to be company leadership requesting an immediate payment.
- Audio Message: A voicemail or phone call uses a cloned voice of a trusted colleague or superior giving payment authorization.
- Combination Attack: A multi-channel approach uses fake emails, followed by a deepfake voice call to add credibility to a fraudulent payment request.
Real-World Examples
The German Energy Company Scam (2019) – In what is considered the first case of its kind, fraudsters used an artificial intelligence (AI) voice clone of a German energy boss to scam the head of a UK subsidiary out of €220,000. This early case demonstrated the potential for voice cloning technology to bypass traditional security measures.
The $25 Million Hong Kong Heist (2024) – As previously detailed by VendorInfo, in February 2024, scammers impersonating a company executive used AI-generated voice to trick a worker at Arup Engineering into transferring $25 million. The voice was flawless. The video call was populated with realistic deepfake avatars. The employee participated in what appeared to be a legitimate video conference with multiple executives, all of whom were deepfake impersonations. The scam involving the fake CFO was only discovered when the employee later checked with the corporation’s head office.
Internal YouTube Warning (2025) – In March, YouTube warned employees about a privately circulated AI-generated video presenting a very convincing image of CEO Neal Mohan.
Rising Frequency and Impact
According to research from Starling Bank, published in September, Business Email Compromise (BEC) and vishing scams caused $2.6 billion in losses globally, with deepfake fraud incidents rising tenfold from 2022 to 2023.
In 2024, businesses had an average loss of nearly $500,000 due to deepfake-related fraud, with large enterprises experiencing losses up to $680,000.
According to a recent article in the Wall Street Journal[2], there were more than 105,000 deepfake-related attacks in the U.S. last year. “Roughly an attack every five minutes—a massive jump from 2023,” said Brian Long, chief executive and co-founder of OpenAI-backed cybersecurity firm Adaptive Security.
“A year ago, maybe one in 10 security executives I spoke to had seen one. Now it’s closer to five in 10,” said Long, whose firm specializes in guarding against AI-powered social engineering attacks.
The Journal story goes on to report attacks on Ferrari, cloud security company Wiz, and the advertising firm WPP.
So beware the growing threat of deep fake impersonations and develop a plan to protect your organization. See a list of red flags and what to do in the article Protect Against Deepfake Impersonation Fraud in Accounts Payable- Part II.
To learn how VendorInfo’s bank account verification tool can be a vital part of your fraud protection protocols, contact us.
[1] https://malwaretips.com/blogs/pink-salt-trick-for-weight-loss/
[2] https://www.wsj.com/articles/ai-drives-rise-in-ceo-impersonator-scams-2bd675c4?mod=djemCybersecruityPro&tpl=cs


