North Korea-Linked Actors Allegedly Use AI Executive Deepfakes in Zoom Phishing Targeting Web3 Employee

An alleged phishing scheme involving actors linked to North Korea used purported AI-generated deepfake videos of company executives to deceive a Web3 employee during a fake Zoom call. The target was reportedly tricked into installing macOS malware disguised as a "Zoom extension," leading to the deployment of spyware, a keylogger, and a crypto wallet stealer. The attackers reportedly used Telegram and spoofed Zoom domains to orchestrate the breach.

Risk classification

• Primary risk domain: 4 Malicious actors
• Primary risk subdomain: 4.3 Fraud, scams, and targeted manipulation

The attackers used AI-generated deepfakes of company executives to impersonate trusted individuals in a targeted Zoom call, aiming to deceive a specific employee into installing malware for financial theft.

Causal factors

• Entity: Human
• Intent: Intentional
• Timing: Post-deployment

The incident was caused by human malicious actors who intentionally deployed AI-generated deepfakes post-deployment to deceive the victim.

EU AI Act risk tier

• Risk tier: 3 Limited Risk

Limited Risk: The report describes the use of AI-generated content such as deepfakes, which are subject to transparency obligations under the EU AI Act.

AI system and alleged parties

• AI system: unspecified
• AI purpose: Deepfake Video Generation; Voice Generation
• Behaviour type: Tool
• Alleged developer: Unknown voice cloning technology developers, Unknown deepfake technology developers
• Alleged deployer: North Korea, Lazarus Group, Government of North Korea, BlueNoroff
• Alleged harmed parties: Zoom, Web3, Unnamed Web3 employee, Truth, National security and intelligence stakeholders, macOS users, Epistemic integrity, Cryptocurrency infrastructure

Harm severity

Severity ratings are based on this scale.

Highest direct severity in any category: Negligible. Severity is scored from Negligible to Catastrophic in each harm category, for harm the reports describe as caused directly or indirectly by the AI system.

• Physical: direct Negligible, indirect Negligible
• Infrastructure: direct Negligible, indirect Negligible
• Property: direct Negligible, indirect Negligible
• Financial: direct Negligible, indirect Negligible
• Environmental: direct Negligible, indirect Negligible
• Malicious content: direct Negligible, indirect Negligible
• Differential treatment: direct Negligible, indirect Negligible
• Civil rights: direct Negligible, indirect Negligible
• Democracy: direct Negligible, indirect Negligible
• Privacy: direct Negligible, indirect Minor
• Psychological: direct Negligible, indirect Negligible
• Epistemic: direct Negligible, indirect Negligible
• Child sexual exploitation and abuse: direct Negligible, indirect Negligible

Privacy

Reported: The report explicitly describes privacy violations resulting from the installation of malware, including a keylogger and credential harvester.

Indirectly caused: The deepfake deception led the victim to install malware that captured clipboard, screen data, and keystrokes, compromising their personal and corporate credentials.

Inferred additional harm: It is highly likely that sensitive corporate credentials, personal passwords, and private keys were compromised, leading to broader unauthorized access to other systems.

People affected

• Occurrences reported: 1
• People reportedly harmed: 1
• People reportedly exposed: 1

Potential causes

Technology

• AI-Generated Deepfakes: Attackers used realistic AI deepfakes of executives to deceive the employee.
• Malicious AppleScript Bypass: AppleScript bypassed user logs and triggered stealth malware installation.
• C2 Infrastructure Mimicry: Command and control traffic was routed through domains mimicking Zoom.

Data Inputs

• Executive Likeness Data: Attackers harvested executive video and audio to train deepfake models.
• Compromised Credentials: Harvesting of browser credentials and cookies facilitated initial access.

Human Factors

• Social Engineering Trust: Employee trusted the visual deepfakes representing company executives.
• Credential Disclosure: Employee provided system password when prompted by the fake support tool.
• Phishing Susceptibility: Employee accepted Telegram messages and fake Calendly invites.

Process and Methods

• Inadequate Verification: Lack of out-of-band verification for executive identities on calls.
• Unsecured Software Install: Installing unverified Zoom extensions from untrusted chat channels.
• Weak Hiring Assessments: Using fake assessments to trick job-seekers into running code.

Information quality

• Classification confidence: High
• Reason for confidence: The report provides clear details about the attack vector, the role of AI-generated deepfakes in the Zoom call, and the specific malware components deployed. The role of the AI is clearly defined as a deception tool in a spear-phishing campaign.
• Ambiguities identified: The specific AI tools or models used to generate the deepfakes are not identified.
• Alternative interpretations: None. The incident is a classic spear-phishing attack enhanced by deepfake technology.

North Korea-linked threat actor BlueNoroff utilized sophisticated AI-generated deepfakes of company executives during a live video call to compromise a Web3 employee and install data-stealing malware. This represents a significant evolution in state-sponsored social engineering capabilities, highlighting the growing national security threat of synthetic media in cyber warfare.

• Overall national security impact: Substantial
• Scope: Unknown
• Primary target: Unknown
• Other affected: Unknown
• Alleged perpetrator: BlueNoroff (North Korea)

Threat characteristics

• Imminence: Long-term. Represents an ongoing strategic capability and evolving threat vector in cyber espionage rather than an active, unresolved crisis.
• Autonomy: Human-controlled. The AI system was used as a tool to generate synthetic video and audio, with human operators directing the social engineering attack.
• Novelty: Evolved capability. Represents a significant advancement of existing social engineering tactics by integrating highly realistic, real-time synthetic media in live interactions.

Impact by dimension

• Physical security: Negligible. No physical systems, critical infrastructure, or human safety were threatened during this cyber incident.
• Information security: Substantial. State-sponsored actor BlueNoroff utilized AI-generated deepfakes of executives in a live video conference to deceive an employee, demonstrating a highly sophisticated, novel threat vector that warrants serious attention from security agencies.
• Sovereignty: Negligible. The incident targeted a private sector Web3 employee and did not impact government operations, electoral systems, or state sovereignty.
• Economic security: Minor. Targeting of Web3 assets by North Korean state-sponsored actors to exfiltrate cryptocurrency represents a minor economic threat, though part of a larger systemic campaign.
• Societal stability: Negligible. The incident did not involve mass surveillance, civil liberties violations, or threats to societal stability.