This is a research prototype. The data and analyses are preliminary and not yet validated — we'd welcome your .
A job applicant in Tokyo used AI-generated deepfake video and audio to impersonate a real IT executive, Kenbun Yoshii, during a remote hiring interview. The incident is part of a broader trend of suspected North Korean IT workers using AI-generated identities to secure remote employment and generate foreign currency. Security experts and analysis from firms like Okta and Trend Micro confirmed the use of deepfake technology, highlighting the growing risk to corporate hiring processes.
In Tokyo, a Japanese IT company reportedly interviewed a job applicant who used purportedly AI-generated video manipulation to impersonate real IT executive Kenbun Yoshii during a remote hiring interview. Investigators cited visual and audio irregularities suggesting a deepfake, and Yoshii said his publicly available images and career details appeared to have been misused.
Domain classification, causal taxonomy, severity scores, and national security assessments were LLM-classified and may contain errors.
Using AI systems to gain a personal advantage over others such as through cheating, fraud, scams, blackmail or targeted manipulation of beliefs or behavior. Examples include AI-facilitated plagiarism for research or education, impersonating a trusted or fake individual for illegitimate financial benefit, or creating humiliating or sexual imagery.
Also classified as
The fraudulent employment attempts are suspected to be part of a broader campaign to generate foreign currency for North Korea's weapons programs.
The applicant used AI deepfake technology to impersonate a real individual for personal advantage and fraudulent employment.
Human
Due to a decision or action made by humans
Intentional
Due to an expected outcome from pursuing a goal
Post-deployment
Occurring after the AI model has been trained and deployed
The incident was caused by human actors intentionally using a deployed AI deepfake tool to commit identity fraud during a job interview.
A suspected North Korean operative used real-time AI deepfake technology to impersonate a real IT executive during a remote interview in Japan. While this specific attempt failed, the incident highlights an evolved national security threat where hostile state actors leverage generative AI to infiltrate foreign companies, bypass sanctions, and secure funding for state weapons programs.
The AI system acted as a tool to alter facial features and personal credentials in real-time during a remote interview, performing a specific function based on direct input.
Limited Risk: The system is a deepfake generator used to alter video and audio. The EU AI Act classifies deepfakes and AI-generated content under Limited Risk, requiring transparency so users know they are interacting with AI.
The AI system was used to alter the applicant's facial features and voice in real-time to impersonate a real individual during an online interview.
The reports from multiple sources provide consistent details about the specific incident on March 4, 2026, including the name of the impersonated individual, the company's reaction, and technical analysis of the deepfake by Okta and Trend Micro.
None
Technology
Advanced Deepfake Generation
AI synthesized realistic video and audio of the victim to bypass visual checks.
Real-Time Facial Alteration
AI altered facial features in real-time during the live online interview.
Mismatched Audio Lip Sync
Generative AI failed to perfectly sync lip movements with spoken Japanese audio.
Data Inputs
Publicly Available Media
Attackers used victim's public images and videos to seed the deepfake model.
Stolen Career Credentials
Victim's real professional credentials were used to build a fake resume.
Human Factors
Overreliance on Visual Feed
Interviewers initially trusted the video feed without secondary verification.
Process and Methods
Inadequate Identity Verification
The company lacked robust multi-factor authentication for applicants.
Lack of Deepfake Detection Tools
The firm lacked real-time deepfake detection tools during the interview.
Fully Remote Hiring Process
Hiring was conducted entirely online without any in-person verification.
Management
Inadequate AI Risk Assessment
Management failed to assess risks of AI-driven identity spoofing in hiring.
Regulatory Environment
Weak Remote Hiring Regulations
No strict regulatory standards exist for identity verification in remote hiring.
AI system involved
An AI system was involved, but the reports do not identify it.
Identified by automated extraction; can differ from the alleged developers above.
