A wave of sophisticated cyberattacks is leveraging artificial intelligence to impersonate corporate leaders in live video meetings. According to coverage by Escudo Digital based on findings from cybersecurity firm Arctic Wolf, cybercriminals are cloning the faces and voices of founders and senior executives to conduct fraudulent video calls on platforms like Zoom and Microsoft Teams. This evolving threat highlights a critical reality for businesses across the UAE and the broader MENA region: as digital communication becomes central to leadership and business development, verifying the authenticity of video interactions is no longer optional.
For founders, investors, and leadership teams operating in fast-growing regional markets, digital interactions often form the foundation of major partnerships and capital investments. However, as generative tools lower the barrier for creating hyper-realistic synthetic media, reliance on visual inspection during video meetings is no longer sufficient. Establishing rigorous verification protocols and securing organizational communications must become a priority for enterprises across all commercial sectors.
How Cybercriminals Clone Founders for Video Calls
According to Escudo Digital, the recent campaign has been traced back to BlueNoroff, a financial theft cell operating within the North Korean Lazarus group. Arctic Wolf Labs investigators uncovered the operation after gaining access to the attackers' internal archive of video assets and lures. Forensic analysis revealed 349 modified videos, confirming that synthetic media was being deployed to deceive targets primarily in the technology, Web3, and venture capital sectors.
To build these digital replicas, attackers gather publicly available media, including YouTube interviews and public conference presentations. They then process this audio and video through a chain of commercial generative artificial intelligence platforms, including GPT-4o, Grok Imagine, Google Generative AI, and Wondershare AIGC. Security researchers at Arctic Wolf Labs warned that the resulting deepfakes can no longer be distinguished from authentic webcam feeds through simple visual observation. The global scope of the operation has expanded rapidly, growing from just over 400 targets across 20 countries in April 2026 to 677 individuals across 58 countries, with 84% of targets identified as company founders and high-ranking business executives.
The Feedback Loop and Malicious Infrastructure
The attack mechanism functions as a self-sustaining feedback loop where each compromised interaction provides material for the next victim. During a fake video call, attackers often feign technical difficulties. While the target attempts to troubleshoot the simulated video or audio failure, the cybercriminals secretly record the live webcam stream. These freshly captured authentic images are subsequently combined with AI-generated clones to target other contacts within the executive's professional network.
In addition to fake video calls, the campaign incorporates extensive web infrastructure designed to fit into standard executive workflows. Investigators identified 98 malicious domains, 73 of which were newly created, designed to mimic encrypted financial documents, security audit reports, cryptocurrency wallet access portals, Chrome PDF viewer updates, and a media player named NovaStream. This multi-layered approach demonstrates how threat actors combine automated media generation with familiar corporate lures to breach secure organizations.
Protecting Communications and Digital Brand Presence
For organizations operating in competitive regional markets, protecting corporate brand identity and executive communication channels requires both operational discipline and robust public messaging strategies. Working with an experienced digital marketing agency dubai can help brands maintain clear, verified public communication channels and authoritative online presences. However, external reputation management must be paired with strict internal cybersecurity safeguards.
Despite the qualitative leap in media quality enabled by artificial intelligence, researchers emphasize that the underlying intrusion mechanism has not fundamentally changed. The initial deepfake video call serves primarily as a social engineering hook. The ultimate compromise still requires the targeted executive to manually download and execute malicious code, such as installing a fake software patch or running a local command. Consequently, if an executive refuses to run unauthorized software and verifies unexpected requests through an alternative channel, the attack chain is broken entirely.
Key Safeguards for MENA Organizations
- Implement out-of-band verification protocols: Establish mandatory secondary communication channels, such as a pre-agreed phone call or verified messaging route, to confirm identity before acting on urgent financial or operational requests made during video calls.
- Enforce strict endpoint security controls: Restrict administrative privileges on corporate devices to prevent executives and staff from manually executing unapproved local scripts, installing unverified browser extensions, or applying third-party patches.
- Audit and secure official media assets: Regularly review corporate media published on public platforms, ensuring that executive video content is monitored and that public communications reflect official corporate verification standards.
- Train executive teams on social engineering tactics: Conduct specialized awareness sessions for senior leadership, highlighting indicators such as unexpected software installation prompts during video calls or unusual technical failure claims.
- Establish incident response procedures for digital impersonation: Formulate clear protocols for reporting suspected executive impersonation across corporate communication tools, enabling internal IT teams to immediately block related malicious domains.
Frequently asked questions
According to security reports, who is behind these deepfake attacks?
According to cybersecurity firm Arctic Wolf as reported by Escudo Digital, the campaign is linked to BlueNoroff, a financial theft cell integrated into the North Korean cybercrime group known as Lazarus.
How do attackers obtain the video and voice data needed for deepfakes?
Attackers extract public video material, such as YouTube interviews and conference talks, and process them through commercial generative artificial intelligence tools including GPT-4o, Grok Imagine, Google Generative AI, and Wondershare AIGC.
What halts the deepfake intrusion process before harm occurs?
Although AI acts as a production multiplier for producing realistic video replicas, the final attack phase relies on executives manually executing malicious code. The entire intrusion stops if an executive distrusts urgent requests to install software or run local commands and verifies the request through an alternative communication channel.
Maintaining resilient communication channels and a trusted digital presence is essential for business growth across the region. To discuss how to protect your brand messaging and establish multi-channel communication strategies, message us on WhatsApp or book a call with Extreme Marketing today.
Sources and further reading
Want help putting this into practice?
Book a free audit and we will review your marketing and suggest the quickest wins.
