Choose your location

Latest
Thomas Markle charged over child sex abuse material in Philippines University of Birmingham pauses plans to move Shakespeare Institute after outcry Government backs welfare shake-up to prevent 'lost generation' CSJMU prominence grows in QS Sustainability Rankings, grabs top spot among state univs Cop31: UK backs idea of new levies on polluting sectors to pay for climate action costs Scheifele praises Hawerchuk after breaking his Jets' points record: 'I wouldn't be where I am without him'

North Korean hackers clone CEOs' faces and voices for deepfake Zoom scams

North Korean cybercriminals are extracting YouTube interviews and conference talks to use artificial intelligence to clone the faces and voices of company founders and CEOs with the aim of impersonating them...

Escudo Digital Alberto Payo

North Korean cybercriminals are extracting YouTube interviews and conference talks to use artificial intelligence to clone the faces and voices of company founders and CEOs with the aim of impersonating them in fake Zoom and Microsoft Teams video calls. According to the American cybersecurity firm Arctic Wolf, this sophisticated social engineering scheme has targeted 677 individuals across 58 countries, of which 84% are precisely founders and high-ranking business executives.

The discovery was made possible because Arctic Wolf Labs investigators managed to access the attackers' internal repository of videos and lures, linked to BlueNoroff—a financial theft cell integrated into the well-known Lazarus group from North Korea . In the forensic analysis of this private archive, analysts counted 349 videos marked as probably modified , confirming through digital provenance tests the existence of dozens of deepfakes and manipulated material created to deceive executives in the tech, Web3, and venture capital sectors.

To create these digital replicas, the attackers have abandoned rudimentary methods and use a chain of commercial generative artificial intelligence tools like GPT-4o, Grok Imagine, Google Generative AI, and Wondershare AIGC. "The videos they produce can no longer be distinguished from real webcam images through simple visual inspection," warn Arctic Wolf Labs about the evolution of the campaign in the past six months.

One deception leads to another The system works as a continuous feedback loop where each new victim feeds the next deception. When an executive falls into the trap during a supposed meeting with known partners or investors, the cybercriminals secretly record the live transmission of their webcam while they try to solve a simulated technical failure, later reusing those real images along with the AI-generated clones to deceive other contacts in their professional circle .

Beyond the fictitious video calls, the group has diversified its lures into the daily workflow of senior management, designing 98 malicious domains (73 of them new) that mimic encrypted financial documents, security audits, access to crypto wallets, or fake updates of Chrome's PDF viewer and a player called NovaStream.

In total, the number of identified targets has grown from just over 400 recorded in 20 countries in April 2026 to the current 677. Despite the qualitative leap that AI automation represents, researchers emphasize that t he final phase of the attack still depends on the executive manually executing the malicious code.

"Artificial intelligence now acts as a production multiplier that helps BlueNoroff scale and diversify, but it has not fundamentally changed the intrusion mechanism," Arctic Wolf notes, reminding that the whole operation stops if the executive distrusts urgent requests to install patches or execute local commands and verifies the request through an alternative means.

North Korean cybercriminals are extracting YouTube interviews and conference talks to use artificial intelligence to clone the faces and voices of company founders and CEOs with the aim of impersonating them in fake Zoom and Microsoft Teams video calls. According to the American cybersecurity firm Arctic Wolf, this sophisticated social engineering scheme has targeted 677 individuals across 58 countries, of which 84% are precisely founders and high-ranking business executives.

The discovery was made possible because Arctic Wolf Labs investigators managed to access the attackers' internal repository of videos and lures, linked to BlueNoroff—a financial theft cell integrated into the well-known Lazarus group from North Korea . In the forensic analysis of this private archive, analysts counted 349 videos marked as probably modified , confirming through digital provenance tests the existence of dozens of deepfakes and manipulated material created to deceive executives in the tech, Web3, and venture capital sectors.

To create these digital replicas, the attackers have abandoned rudimentary methods and use a chain of commercial generative artificial intelligence tools like GPT-4o, Grok Imagine, Google Generative AI, and Wondershare AIGC. "The videos they produce can no longer be distinguished from real webcam images through simple visual inspection," warn Arctic Wolf Labs about the evolution of the campaign in the past six months.

One deception leads to another The system works as a continuous feedback loop where each new victim feeds the next deception. When an executive falls into the trap during a supposed meeting with known partners or investors, the cybercriminals secretly record the live transmission of their webcam while they try to solve a simulated technical failure, later reusing those real images along with the AI-generated clones to deceive other contacts in their professional circle .

Beyond the fictitious video calls, the group has diversified its lures into the daily workflow of senior management, designing 98 malicious domains (73 of them new) that mimic encrypted financial documents, security audits, access to crypto wallets, or fake updates of Chrome's PDF viewer and a player called NovaStream.

In total, the number of identified targets has grown from just over 400 recorded in 20 countries in April 2026 to the current 677. Despite the qualitative leap that AI automation represents, researchers emphasize that t he final phase of the attack still depends on the executive manually executing the malicious code.

"Artificial intelligence now acts as a production multiplier that helps BlueNoroff scale and diversify, but it has not fundamentally changed the intrusion mechanism," Arctic Wolf notes, reminding that the whole operation stops if the executive distrusts urgent requests to install patches or execute local commands and verifies the request through an alternative means.

Article text via FreeNewsAPI. Rights remain with Escudo Digital.

Read on publisher site → Opens Escudo Digital in a new tab

More in Technology