Within Deepfakes
Seeing Them on a Call Is Not Enough
A video meeting can feel socially real, but unusual financial requests still need independent approval channels.
On this page
- How fabricated meetings create false authority
- Payment red flags inside urgent workplace requests
- Approval habits that survive convincing video
Page outline Jump by section
Introduction
Seeing a familiar face on a video call feels like strong evidence. In workplaces, that feeling can be dangerous. Deepfake video-call fraud exploits a simple human shortcut: people tend to trust what appears to be a live conversation, especially when it involves recognised colleagues, senior managers, or financial decision-makers. Recent cases show that criminals can use AI-generated video and audio to create convincing meetings in which fake executives appear to authorise transfers, approve payments, or demand urgent action. The critical lesson is that payment approval should never depend solely on what appears on a screen. Financial verification must rely on independent procedures that remain valid even when video itself cannot be trusted. [Adaptive Security]adaptivesecurity.comAdaptive Security Arup Deepfake Scam: How $25M Was Stolen via Video CallAdaptive SecurityArup Deepfake Scam: How $25M Was Stolen via Video CallMay 16, 2024 — Explore how engineering firm Arup lost $25 million…
How Fabricated Meetings Create False Authority
Deepfake payment fraud is not primarily a technical attack on software. It is a social attack on trust.
Traditional business email compromise schemes often relied on fake emails from a supposed chief executive or finance director. Deepfake video calls extend the same strategy into a more persuasive environment. Instead of reading a suspicious message, employees believe they are participating in a live meeting with familiar people. The visual presence of managers, colleagues and corporate branding can create a powerful sense of legitimacy. [Internet Crime Complaint Center]ic3.govInternet Crime Complaint CenterBusiness Email Compromise: Virtual Meeting PlatformsFeb 16, 2022 — The scam is frequently carried out when…
The best-known example emerged in Hong Kong in 2024. A finance employee attended a video conference that appeared to include the company’s chief financial officer and several colleagues. After the meeting, the employee completed a series of transfers totalling roughly US$25 million. Investigators later concluded that every participant other than the victim had been generated or manipulated using AI. The employee’s initial doubts were overcome because multiple apparently trusted figures on the call reinforced the request. [Financial Times+2Financial Times]ft.comFinancial TimesArup lost $25mn in Hong Kong deepfake video conference…16 May 2024 — UK engineering group Arup lost HK$200mn ($25mn) af…
What makes these incidents notable is that the deception does not depend on a perfect digital imitation. The fraud succeeds because several trust signals operate simultaneously:
- A familiar face appears on screen.
- The request comes from someone higher in the organisational hierarchy.
- Multiple participants seem to confirm the instruction.
- The meeting format feels normal and routine.
- The transaction is framed as confidential or time-sensitive.
Each signal may be weak on its own. Together they create a false impression that verification has already occurred. [McAfee]mcafee.comMc Afee How Scammers Used Deepfake Video to Dupe a CompanyHow Scammers Used Deepfake Video to Dupe a Company…February 27, 2026 — A $25 million deepfake scam: How scammers used AI in a vi…
Why Seeing Several Colleagues Can Be More Convincing Than Seeing One
Many people assume a deepfake attack involves a single impersonated executive. The Hong Kong case demonstrated a more sophisticated approach: a group setting.
In ordinary workplace decision-making, employees often look to other participants for reassurance. If several familiar faces appear to agree with a payment request, the employee may interpret that agreement as evidence that due diligence has already been performed elsewhere. Criminals exploit this tendency by creating a simulated consensus. The victim is not merely trusting a face; they are trusting apparent group validation. [Adaptive Security+2McAfee]adaptivesecurity.comAdaptive Security Arup Deepfake Scam: How $25M Was Stolen via Video CallAdaptive SecurityArup Deepfake Scam: How $25M Was Stolen via Video CallMay 16, 2024 — Explore how engineering firm Arup lost $25 million…
This is one reason why “I saw them in a meeting” is no longer a sufficient verification standard for financial approvals.
Payment Red Flags Inside Urgent Workplace Requests
Deepfake-enabled fraud usually focuses on changing behaviour rather than proving identity. The most important warning signs are often found in the request itself.
A payment instruction deserves additional scrutiny when it combines authority with urgency. Criminals frequently attempt to create conditions in which normal checks feel inconvenient or inappropriate. FBI warnings about AI-assisted fraud emphasise that attackers increasingly use convincing voice and video impersonation to pressure individuals and businesses into acting quickly. [FBI]fbi.govwarns of increasing threat of cyber criminals utilizing artificial intelligenceFBI Warns of Increasing Threat of Cyber Criminals Utilizing…8 May 2024 — Attackers are leveraging AI to craft highly convincing voi…
Common red flags include:
- Requests to bypass standard approval procedures.
- Claims that a transaction is highly confidential.
- Pressure to act before a deadline or market event.
- Instructions to use a new bank account or unfamiliar payment route.
- Resistance when independent confirmation is suggested.
- Claims that normal approvers are unavailable.
- Multiple rapid transfers rather than a single documented payment.
The Hong Kong incident reportedly involved numerous transfers to several accounts rather than a single routine payment. That pattern itself should have triggered heightened verification procedures. [Adaptive Security]adaptivesecurity.comAdaptive Security Arup Deepfake Scam: How $25M Was Stolen via Video CallAdaptive SecurityArup Deepfake Scam: How $25M Was Stolen via Video CallMay 16, 2024 — Explore how engineering firm Arup lost $25 million…
Another warning sign is procedural inconsistency. A genuine executive may occasionally request urgent action, but legitimate organisations usually have documented approval processes precisely for high-value transactions. When a request requires abandoning those controls, the risk comes not from the technology used on the call but from the attempt to override established safeguards. [Internet Crime Complaint Center]ic3.govInternet Crime Complaint CenterBusiness Email Compromise: Virtual Meeting PlatformsFeb 16, 2022 — The scam is frequently carried out when…
Approval Habits That Survive Convincing Video
The most reliable defence against deepfake video fraud is not perfect detection technology. It is independent verification.
Research and law-enforcement guidance increasingly point toward a decision-layer approach: assume that video, audio, email and chat messages can potentially be manipulated, then build approval processes that do not rely on any single communication channel. [arXiv]arxiv.orgSynthetic Trust Attacks: Modeling How Generative AI Manipulates Human Decisions in Social Engineering FraudApril 3, 2026…
Practical approval habits include:
Separate communication from authorisation.
A video call can initiate a discussion, but payment approval should occur through established financial controls rather than through the meeting itself.
Use a second channel for confirmation.
If a transfer request arrives through a video conference, confirm it through a known telephone number, internal workflow system, or another pre-approved channel. The key is independence from the original communication. [FBI]fbi.govwarns of increasing threat of cyber criminals utilizing artificial intelligenceFBI Warns of Increasing Threat of Cyber Criminals Utilizing…8 May 2024 — Attackers are leveraging AI to craft highly convincing voi…
Require multiple approvers for significant payments.
Dual-control and multi-signature procedures reduce the chance that one manipulated employee can authorise a fraudulent transaction.
Treat urgency as a reason to slow down.
The more pressure attached to a payment request, the more valuable independent verification becomes.
Verify account changes separately.
New banking details should be confirmed through established contact information rather than information supplied during the same meeting.
These habits remain effective whether the fraud uses email, voice cloning, video deepfakes or future forms of synthetic media because they focus on the decision rather than the appearance of authenticity. [arXiv]arxiv.orgSynthetic Trust Attacks: Modeling How Generative AI Manipulates Human Decisions in Social Engineering FraudApril 3, 2026…
Why Detection Alone Is Not Enough
A common reaction to deepfake fraud is to look for visual clues: unnatural blinking, distorted mouths, mismatched lighting or audio glitches. Such indicators can sometimes reveal manipulation, but relying on them creates a fragile defence.
Recent research shows that deepfake detection systems often perform substantially worse on real-world material than on controlled test datasets. Academic evaluations have found significant drops in accuracy when detectors encounter newer, in-the-wild deepfakes rather than benchmark examples. [arXiv]arxiv.orgDeepfake-Eval-2024: A Multi-Modal In-the-Wild Benchmark of Deepfakes Circulated in 2024March 4, 2025…
This matters because workplace payment decisions occur under time pressure. Employees cannot realistically perform forensic analysis during a live meeting. Even if technical detection tools improve, organisations still need approval processes that assume some fraudulent communications will appear convincing.
The deeper lesson is that credibility and authenticity are not the same thing. A video call may appear authentic while still being part of a deception. Payment verification should therefore depend on documented procedures, independent confirmation and accountable approval chains rather than on the apparent realism of a face on a screen. The safest mindset is simple: seeing someone on a call may justify further discussion, but it should never be the final proof required to move money. [arXiv+2FBI]arxiv.orgSynthetic Trust Attacks: Modeling How Generative AI Manipulates Human Decisions in Social Engineering FraudApril 3, 2026…
Amazon book picks
Further Reading
Books and field guides related to Seeing Them on a Call Is Not Enough. Use these as the next step if you want deeper reading beyond the article.
Thinking, Fast and Slow
Helps readers understand cognitive shortcuts that fraudsters exploit.
The Art of Deception
First published 2001. Subjects: Computer Technology, Nonfiction, Social aspects, Computer hackers, Computer security.
Social Engineering
First published 2011. Subjects: Computer hackers, Social engineering, Computer security, Human-computer interaction, Hackers.
eBay marketplace picks
Marketplace Samples
Example marketplace items related to this page. Use the search link to explore similar finds on eBay.
![Listing image for NCT 127 [FACT CHECK] 5th Album CHANDELIER Ver/CD+Photo Book+Photo Card+Poster](/assets/images/marketplace-covers/0ec43911bad3ef719bae.jpg)
Endnotes
-
Source: fbi.gov
Title: warns of increasing threat of cyber criminals utilizing artificial intelligence
Link: https://www.fbi.gov/contact-us/field-offices/sanfrancisco/news/fbi-warns-of-increasing-threat-of-cyber-criminals-utilizing-artificial-intelligenceSource snippet
FBI Warns of Increasing Threat of Cyber Criminals Utilizing...8 May 2024 — Attackers are leveraging AI to craft highly convincing voi...
Published: May 2024
-
Source: mcafee.com
Title: Mc Afee How Scammers Used Deepfake Video to Dupe a Company
Link: https://www.mcafee.com/blogs/ai/how-scammers-used-deepfake-video-to-dupe-a-company-out-of-millions/Source snippet
How Scammers Used Deepfake Video to Dupe a Company...February 27, 2026 — A $25 million deepfake scam: How scammers used AI in a vi...
Published: February 27, 2026
-
Source: arxiv.org
Link: https://arxiv.org/abs/2604.04951Source snippet
Synthetic Trust Attacks: Modeling How Generative AI Manipulates Human Decisions in Social Engineering FraudApril 3, 2026...
Published: April 3, 2026
-
Source: arxiv.org
Link: https://arxiv.org/abs/2503.02857Source snippet
Deepfake-Eval-2024: A Multi-Modal In-the-Wild Benchmark of Deepfakes Circulated in 2024March 4, 2025...
Published: March 4, 2025
-
Source: fbi.gov
Link: https://www.fbi.gov/video-repository/psa-business-e-mail-compromise-scam.mp4/view -
Source: fbi.gov
Link: https://www.fbi.gov/investigate/cyber -
Source: fbi.gov
Link: https://www.fbi.gov/how-we-can-help-you/scams-and-safety/common-frauds-and-scams/business-email-compromise -
Source: adaptivesecurity.com
Title: Adaptive Security Arup Deepfake Scam: How $25M Was Stolen via Video Call
Link: https://www.adaptivesecurity.com/blog/arup-deepfake-scam-attackSource snippet
Adaptive SecurityArup Deepfake Scam: How $25M Was Stolen via Video CallMay 16, 2024 — Explore how engineering firm Arup lost $25 million...
Published: May 16, 2024
-
Source: ic3.gov
Link: https://www.ic3.gov/PSA/2022/PSA220216Source snippet
Internet Crime Complaint CenterBusiness Email Compromise: Virtual Meeting PlatformsFeb 16, 2022 — The scam is frequently carried out when...
-
Source: ft.com
Link: https://www.ft.com/content/b977e8d4-664c-4ae4-8a8e-eb93bdf785ea?syn-25a6b1a6=1Source snippet
Financial TimesArup lost $25mn in Hong Kong deepfake video conference...16 May 2024 — UK engineering group Arup lost HK$200mn ($25mn) af...
Published: May 2024
-
Source: ft.com
Link: https://www.ft.com/content/b977e8d4-664c-4ae4-8a8e-eb93bdf785eaSource snippet
Fraudsters used a digitally cloned version of a senior manager during a video conference to order financial transfers. This incident, one...
-
Source: ic3.gov
Link: https://www.ic3.gov/PSA/2024/PSA241203Source snippet
Internet Crime Complaint CenterCriminals Use Generative Artificial Intelligence to Facilitate...3 Dec 2024 — The FBI is warning the publ...
-
Source: ic3.gov
Link: https://www.ic3.gov/PSA/2025/PSA250515Source snippet
Internet Crime Complaint CenterSenior US Officials Impersonated in Malicious Messaging...May 15, 2025 — Malicious actors have impersonat...
Published: May 15, 2025
-
Source: arup.com
Title: Planning, designing, engineering the future
Link: https://www.arup.com/Source snippet
Planning, designing, engineering the future - ArupA global consultancy providing expertise across 150+ disciplines: shaping infras...
-
Source: counterfraud.gov.au
Link: https://www.counterfraud.gov.au/case-studies/company-worker-hong-kong-pays-out-ps20m-deepfake-video-call-scamSource snippet
m) of her firm's money to fraudsters in a deepfake video conference call.Read more...
-
Source: ibm.com
Title: business email compromise
Link: https://www.ibm.com/think/topics/business-email-compromiseSource snippet
(BEC)In early 2023, the FBI warned of a new type of attack in which scammers pose as corporate customers to steal products from the targe...
Additional References
-
Source: cdn.lawreportgroup.com
Link: https://cdn.lawreportgroup.com/acuris/files/ACR-New/AI%20Incidents%20%E2%80%A2%20Incident%20634_%20Alleged%20Deepfake%20CFO%20Scam%20Reportedly%20Costs%20Multinational%20Engineering%20Firm%20Arup%20%2425%20Million.pdfSource snippet
634: Alleged Deepfake CFO Scam Reportedly...Description: A finance employee at the multinational engineering firm Arup was reportedly de...
-
Source: abnormal.ai
Link: https://abnormal.ai/glossary/ceo-fraudSource snippet
What is CEO Fraud? How to Identify & Stop ItCEO fraud is a phishing campaign using executive impersonation or an account takeover to tric...
-
Source: linkedin.com
Link: https://www.linkedin.com/posts/anderssormannilsson_cybersecurity-deepfake-fraudprevention-activity-7459361968102088704-DWv7Source snippet
Anders Sorman-Nilsson posted on the topicAn employee was nearly tricked into transferring funds to 11 Hong Kong bank accounts after a Zoo...
-
Source: linkedin.com
Link: https://www.linkedin.com/posts/ashermanxv_the-fbi-recently-issued-a-stark-warning-activity-7331706738917761024-hpXLSource snippet
Andrey Sherman's PostThe FBI recently issued a stark warning: cybercriminals are using AI-generated voice deepfakes to impersonate U.S. o...
-
Source: reuters.com
Link: https://www.reuters.com/legal/legalindustry/real-insurance-coverage-increasing-ai-deepfake-risks-2024-04-11/Source snippet
Deepfakes use AI to create realistic but false representations, such as altering facial features or mimicking voices, leading to scams th...
-
Source: theguardian.com
Link: https://www.theguardian.com/world/2024/feb/05/hong-kong-company-deepfake-video-conference-call-scamSource snippet
La policía de Hong Kong recibió el informe el 29 de enero, y clasificó el caso como "obtainación de propiedad mediante engaño", siendo ma...
-
Source: vendorinfo.com
Link: https://vendorinfo.com/deepfake-a-horrifying-tale-of-a-25-million-cybercrime/Source snippet
Catherine of Siena. Neither do cybercriminals. And with the arrival of generative AI and deep fakes, the newest cyber...Read more...
-
Source: elite.nz
Title: the deepfake ceo scam why voice cloning is the new business email compromise bec
Link: https://elite.nz/the-deepfake-ceo-scam-why-voice-cloning-is-the-new-business-email-compromise-bec/Source snippet
The “Deepfake CEO” Scam: Why Voice Cloning Is the New...15 Feb 2026 — “Vishing” (voice phishing) uses AI voice cloning to bypass the var...
-
Source: brside.com
Title: what is deepfake bec how voice cloning replaced the wire transfer email
Link: https://www.brside.com/blog/what-is-deepfake-bec-how-voice-cloning-replaced-the-wire-transfer-emailSource snippet
What Is Deepfake BEC? How Voice Cloning Replaced...1 Apr 2026 — Deepfake BEC (Business Email Compromise) is a fraud attack where crimina...
-
Source: linkedin.com
Title: deepfake executives real money how arup tricked us25m shamsuddin 1smnc
Link: https://www.linkedin.com/pulse/deepfake-executives-real-money-how-arup-tricked-us25m-shamsuddin-1smncSource snippet
Deepfake Executives, Real Money: How Arup Was Tricked...In 2024, global engineering firm Arup confirmed it was the victim of a deepfake...
Topic Tree