Chinese Province Developed System Tracking Journalists and International Students

Henan's provincial government reportedly planned system involving facial recognition cameras connected to regional and national databases specifically to track foreign journalists and international students.

Risk classification

• Primary risk domain: 4 Malicious actors
• Primary risk subdomain: 4.1 Disinformation, surveillance, and influence at scale

The Henan government commissioned an AI-powered surveillance system to conduct large-scale, targeted tracking of journalists, international students, and Uyghurs, which constitutes malicious surveillance at scale.

Additional risk subdomains

• 2.1 Compromise of privacy by obtaining, leaking or correctly inferring sensitive information: The system uses facial recognition and database integration to compile private files on individuals without their consent.
• 1.1 Unfair discrimination and misrepresentation: The system explicitly targets and tracks individuals based on their ethnicity (Uyghurs) and nationality.

Causal factors

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

The risk is caused by the intentional decision of human actors (the Henan government and Neusoft) to design and deploy a targeted surveillance system.

EU AI Act risk tier

• Risk tier: 1 Unacceptable

Unacceptable Risk: The system represents a government-run surveillance and social categorization system that grades individuals (journalists) into risk categories (red, yellow, green) to restrict their rights, which constitutes prohibited social scoring and unacceptable biometric surveillance.

AI system and alleged parties

• AI system: FusionInsight, Henan surveillance system (Huawei)
• AI purpose: Smart Surveillance; Face Recognition
• Behaviour type: Autonomous
• Alleged developer: Neusoft
• Alleged deployer: Henan Public Security Department, Henan government
• Alleged harmed parties: international students in Henan, foreign journalists in Henan

Harm severity

Severity ratings are based on this scale.

Highest direct severity in any category: Substantial. 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 Minor, indirect Negligible
• Environmental: direct Negligible, indirect Negligible
• Malicious content: direct Negligible, indirect Negligible
• Differential treatment: direct Substantial, indirect Negligible
• Civil rights: direct Substantial, indirect Negligible
• Democracy: direct Substantial, indirect Negligible
• Privacy: direct Substantial, indirect Negligible
• Psychological: direct Negligible, indirect Negligible
• Epistemic: direct Negligible, indirect Negligible
• Child sexual exploitation and abuse: direct Negligible, indirect Negligible

Financial

Reported: The report describes 1 contract valued at 5 million yuan ($782,000) awarded to Neusoft for the surveillance system.

Directly caused: The average financial cost per occurrence is the contract value of approximately $782,000 spent by the Henan government to procure the system.

Differential treatment

Reported: The report explicitly describes differential treatment of individuals and groups based on their profession, nationality, and ethnicity.

Directly caused: The system is designed to specifically track and categorize journalists into risk levels (red, yellow, green) and monitor international students, migrant women, and Uyghurs.

Inferred additional harm: The system will likely lead to systematic discrimination, increased police harassment, and targeted restrictions on movement for minority groups and journalists in Henan.

Civil rights

Reported: The report explicitly describes plans to use the surveillance system to track and suppress journalists and international students, violating their civil rights.

Directly caused: The system streamlines state suppression of the press by tracking journalists' movements, hotel registrations, and travel to obstruct their work.

Inferred additional harm: The system is highly likely to result in widespread violations of freedom of the press, freedom of movement, and equal protection under the law for thousands of targeted individuals.

Democracy

Reported: The report describes the erosion of press freedom and tightening of media control by the ruling Chinese Communist Party.

Directly caused: The system is designed to quickly locate journalists and obstruct their work, directly undermining press freedom and democratic oversight.

Inferred additional harm: The deployment of such custom suppression technology will likely further entrench authoritarian control and completely eliminate independent journalism in the region.

Privacy

Reported: The report explicitly describes massive privacy violations through the compilation of individual files using facial recognition and database integration.

Directly caused: The system compiles 'one person, one file' records using 3,000 facial recognition cameras and links them to national and regional databases to track individuals' movements and travel bookings.

Inferred additional harm: The system will likely violate the privacy rights of millions of citizens in Henan whose data is processed and matched by the surveillance network without consent.

People affected

• Occurrences reported: 1
• People reportedly exposed: 10000

Potential causes

Management

• Prioritizing Control Over Rights: Government priorities focus on state security and media suppression.
• Corporate Negligence of AI Ethics: Huawei and Neusoft fail to audit how their cloud and AI systems are used.
• Opaque Project Implementation: Removal of public tenders and refusal to respond to inquiries hides risks.

Technology

• Facial Recognition for Masked Faces: Cameras identify individuals even when wearing masks or glasses.
• One Person One File AI Tracking: Huawei AI links real-time video feeds with extensive personal profiles.
• Automated Risk Classification: Algorithmic categorization of journalists into red, yellow, green risk levels.

Data Inputs

• National Database Integration: System connects facial feeds to national and regional databases.
• Multi-Source Personal Data: Uses hotel registrations, travel bookings, and mobile positioning.
• Targeted Group Registries: Maintains libraries of journalists, foreign students, and ethnic minorities.

Human Factors

• State Suppression of Press: Government officials desire to tail and control journalists to limit reporting.
• Lack of Corporate Accountability: Tech firms claim no knowledge or refuse to comment on unethical use.
• Nationalist Online Campaigns: Public harassment of journalists on Weibo drives demand for tracking.

Process and Methods

• Tender and Procurement Process: Rapid procurement of custom surveillance systems with minimal oversight.
• Wartime Early Warning Mechanism: Processes designed to launch heightened security alerts during sensitive dates.
• Lack of Impact Assessments: No human rights impact assessments conducted before deploying AI tracking.

Regulatory Environment

• Absence of Privacy Protections: No legal frameworks to prevent state-sponsored AI surveillance of citizens.
• Standardized Surveillance Systems: National standards like GA/T1400 facilitate massive facial data archiving.
• Lack of International Oversight: International standards fail to stop domestic deployment of repressive AI.

Information quality

• Classification confidence: High
• Reason for confidence: The assessment is based on official government tender documents (190 pages) obtained and archived by IPVM and analyzed by Reuters. The documents explicitly detail the technical requirements, target groups, and contract awards, leaving very little ambiguity about the system's intended capabilities and design.
• Ambiguities identified: The primary ambiguity is whether the system is currently fully operational, as Reuters and IPVM could not independently verify its active deployment status.
• Alternative interpretations: The Chinese government claims such systems are intended to protect national sovereignty and security rather than suppress press freedom, though the explicit targeting of journalists contradicts this framing.

The Henan provincial government's deployment of a custom AI surveillance system to track and categorize journalists, international students, and Uyghurs represents a substantial escalation in automated state-led suppression. While primarily a domestic human rights and societal stability issue, the targeting of foreign journalists and students raises notable concerns for international security and press freedom.

• Overall national security impact: Substantial
• Scope: Multiple nations
• Primary target: China
• Other affected: Nations of foreign journalists and international students
• Alleged perpetrator: Henan Provincial Government

Threat characteristics

• Imminence: Long-term. Represents an ongoing strategic surveillance capability and long-term human rights concern rather than an immediate 72-hour crisis.
• Autonomy: Human-supervised. AI automatically processes video feeds, compiles files, and triggers alerts, but human officials and police officers make final intervention decisions.
• Novelty: Evolved capability. Significant advancement of existing facial recognition and database technologies to automate the tracking of specific professional and ethnic groups.

Impact by dimension

• Physical security: Negligible. No physical threats, kinetic actions, or critical infrastructure compromise reported in connection with this surveillance system.
• Information security: Minor. Involves domestic surveillance and tracking of journalists, but does not represent a compromise of classified intelligence or active foreign information operations.
• Sovereignty: Minor. Represents provincial government exercising domestic authority, though the tracking of foreign nationals could cause minor diplomatic or sovereignty friction.
• Economic security: Minor. Procurement of domestic AI technology from Huawei and Neusoft for 5 million yuan represents local state spending but no major economic disruption.
• Societal stability: Substantial. Substantial impact on human rights and press freedom through automated, systematic tracking and risk-categorization of journalists, international students, and ethnic minorities.