This is a research prototype. The data and analyses are preliminary and not yet validated — we'd welcome your .
The city of Bucheon, South Korea, implemented an AI-powered facial recognition system to track COVID-19 patients and their contacts using thousands of CCTV cameras. While officials argued the system improved contact tracing efficiency, it faced significant criticism from lawmakers and human rights advocates over privacy concerns, lack of public consent, and the potential for long-term data misuse.
Bucheon government’s use of facial recognition in analyzing CCTV footage, despite gaining wide public support, was scrutinized by privacy advocates and some lawmakers for collecting data without consent, and retaining and misusing data beyond pandemic needs.
Domain classification, causal taxonomy, severity scores, and national security assessments were LLM-classified and may contain errors.
AI systems that memorize and leak sensitive personal data or infer private information about individuals without their consent. Unexpected or unauthorized sharing of data and information can compromise user expectation of privacy, assist identity theft, or cause loss of confidential intellectual property.
Also classified as
The system overrides individual consent by tracking non-consenting individuals using silhouettes and clothing, removing their agency over being monitored.
The AI system is designed to track the movements and contacts of individuals using facial recognition and CCTV cameras, directly compromising their privacy and tracking sensitive location data without explicit consent.
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 privacy risks and public tracking are caused by the intentional decision of human government officials to deploy the facial recognition system.
Added categories peak at 2: Minor (Psychological Harm, Inferred)
The deployment of an AI-powered facial recognition and tracking system in Bucheon, South Korea, for COVID-19 contact tracing represents a substantial societal stability and human rights concern. While intended for public health, tracking citizens without consent via a massive CCTV network establishes a concerning precedent for state-level mass surveillance.
The AI system acts as an Assistant designed to digitize manual labor for contact tracers. It analyzes CCTV footage to track movements and contacts, providing this information to human tracing teams to make their work more efficient.
Risk Level 2: High Risk. The system utilizes facial recognition technology for public tracking and surveillance, which falls under high-risk categories such as law enforcement and biometric identification. The report notes it uses 'facial recognition technology to analyse footage gathered by more than 10,820 CCTV cameras and track an infected person's movements'.
The system explicitly uses facial recognition algorithms to identify and track individuals across thousands of CCTV cameras.
The report provides clear, detailed information about the planned pilot project, including the specific technology used (facial recognition and CCTV), the budget (1.6 billion won and 500 million won), the location (Bucheon), and the specific concerns raised by critics. The roles of the AI system and human actors are explicitly defined.
None. The facts of the deployment and the nature of the privacy concerns are straightforward and uncontested, though opinions on whether the privacy trade-off is justified differ.
Technology
Silhouette Tracking Bypass
Tracks individuals via silhouette and clothes even if they refuse consent.
Facial Recognition AI
Uses AI algorithms to analyze CCTV footage and track movements automatically.
Data Inputs
Mass CCTV Footage Integration
Feeds data from over 10,820 city cameras directly into the tracking AI.
Lack of Consent for Data Use
System processes non-consenting individuals' visual data via silhouettes.
Human Factors
Untruthful Patient Testimony
Patients are not always honest about their activities and locations.
Overworked Tracing Teams
Epidemiological investigators working 24-hour shifts created pressure.
Process and Methods
Slow Manual Analysis
Manual CCTV analysis takes hours, driving the push for automated tracking.
Management
Prioritizing Speed Over Privacy
City management prioritized fast tracking over individual privacy concerns.
Regulatory Environment
Infectious Disease Act Loopholes
Legal framework allows invasive tracking, bypassing strict privacy consents.
Lack of Public Consent
System deployed without public consent or explicit data retention limits.
AI system involved
Identified by automated extraction; can differ from the alleged developers above.
