Global AI Surveillance: Every Country That Watches Its Citizens
A country-by-country database of AI-powered surveillance systems: facial recognition, predictive policing, social scoring, and the global infrastructure of algorithmic control.
The Surveillance Explosion
Every continent, every major government, and an increasing number of cities, corporations, and institutions are deploying AI-powered surveillance systems. Facial recognition cameras watch public spaces. Predictive policing algorithms direct law enforcement resources. Social media monitoring tools track political dissent. Biometric databases catalogue populations. And in at least one country, a comprehensive social credit system uses algorithmic assessment to determine citizens’ access to housing, transportation, and employment.
This is not a distant dystopia. It is the present, documented across hundreds of deployments, analyzed by human rights organizations, and expanding faster than any regulatory framework can constrain it. The Carnegie Endowment for International Peace has identified AI surveillance technology in use by at least 75 countries worldwide. The actual number is almost certainly higher, because many deployments are not publicly disclosed.
The global AI surveillance infrastructure is built on a foundation of commercially available technology, much of it developed by a handful of companies and exported with minimal oversight. Understanding which countries use what technology, supplied by whom, and for what purpose is a prerequisite for any meaningful accountability.
China: The Architecture of Total Surveillance
China operates the most extensive AI surveillance infrastructure in the world. It is not close. The system encompasses hundreds of millions of surveillance cameras, comprehensive facial recognition networks, voice recognition systems, gait recognition technology, mobile phone tracking, internet monitoring, and the integration of these data streams into centralized platforms for population management.
The Camera Network
China operates an estimated 600 million or more surveillance cameras, more than any other country by a wide margin. Major cities have achieved near-total coverage of public spaces. The cameras are increasingly equipped with AI capabilities: facial recognition, behavior analysis, crowd monitoring, and anomaly detection.
Facial Recognition
China’s facial recognition systems are deployed across transportation networks, commercial districts, residential compounds, schools, and government buildings. The technology is used for law enforcement (identifying suspects and tracking individuals), access control (building entry, payment systems), and population management (monitoring specific demographic groups, tracking movement patterns).
Chinese facial recognition technology has been developed primarily by domestic companies including Hikvision, Dahua, SenseTime, Megvii, and CloudWalk. These companies have received significant government support and, in several cases, have been implicated in the deployment of surveillance systems targeting Uyghur populations in Xinjiang.
The Xinjiang Model
The surveillance infrastructure deployed in Xinjiang province represents the most intensive application of AI surveillance technology against a civilian population documented anywhere in the world. Reports from human rights organizations, journalists, and leaked internal documents describe a system that includes:
Comprehensive facial recognition scanning at checkpoints, markets, mosques, and other public locations. Mandatory installation of spyware on mobile phones. Predictive policing algorithms that flag individuals for detention based on behavioral indicators such as prayer frequency, foreign contacts, or unusual travel patterns. Biometric data collection including DNA samples, iris scans, and voice prints from entire populations.
The Xinjiang surveillance system is not an aberration. It is a prototype. Technologies and approaches developed and tested in Xinjiang have been deployed, in adapted forms, across China and exported to governments around the world.
Social Credit System
China’s social credit system uses algorithmic assessment to rate individuals, businesses, and government entities based on a range of behavioral and financial indicators. While the system is more fragmented and less monolithic than Western media coverage often suggests, its components include blacklists that restrict access to air and rail travel, public shaming of individuals and companies with low scores, and preferential treatment for those with high ratings.
The social credit system is significant not because it is a single unified platform but because it represents the normalization of algorithmic behavioral assessment at population scale. Its components are being integrated with other surveillance and data systems, creating an increasingly comprehensive infrastructure for monitoring and shaping citizen behavior.
Russia: Surveillance as State Power
Russia has deployed AI surveillance technology primarily through the Moscow municipal government, which has built one of the world’s largest urban facial recognition systems. The Moscow system, developed in partnership with Russian companies including NtechLab, covers the city’s public transportation network, streets, parks, and building entrances.
The system has been used for law enforcement purposes, including identifying suspects and locating wanted individuals. It has also been deployed for political purposes: during protests, facial recognition has been used to identify and subsequently detain participants. Human rights organizations have documented cases in which individuals were detained after being identified by facial recognition at demonstrations.
Russia’s surveillance infrastructure extends beyond Moscow to include internet monitoring through the SORM (System for Operative Investigative Activities) telecommunications interception system, which requires telecom providers to install equipment allowing the Federal Security Service (FSB) direct access to communications data. AI-powered analysis tools have been integrated with SORM to enable automated monitoring of internet communications at scale.
The United States: Fragmented but Extensive
The United States does not have a centralized national surveillance system comparable to China’s. What it has is a fragmented but extensive patchwork of federal, state, and local surveillance capabilities, many of which incorporate AI.
Federal Agencies
The FBI maintains a facial recognition database containing hundreds of millions of photographs, drawn from mugshot databases, visa applications, passport photos, and driver’s license records. The database is used by FBI agents and, through partnerships, by state and local law enforcement agencies.
The Department of Homeland Security operates facial recognition systems at airports and border crossings, comparing travelers’ faces to government databases. Immigration and Customs Enforcement (ICE) has used facial recognition and AI-powered data analysis to identify, locate, and apprehend undocumented immigrants.
The National Security Agency’s signals intelligence capabilities, revealed by Edward Snowden in 2013, include AI-powered analysis of communications metadata and content. The scope and specific capabilities of NSA surveillance remain substantially classified.
State and Local Law Enforcement
Thousands of state and local law enforcement agencies use AI-powered surveillance tools, including facial recognition, predictive policing algorithms, license plate readers, social media monitoring tools, and gunshot detection systems.
Clearview AI, a company that scraped billions of photographs from social media and other public sources to build a facial recognition database, has been used by hundreds of law enforcement agencies across the United States. The company’s practices have been the subject of lawsuits, regulatory actions, and legislative scrutiny, but its technology remains in use.
Predictive policing tools, which use historical crime data to predict where crimes are likely to occur, have been deployed in dozens of US cities. These tools have been criticized for reinforcing existing patterns of racially biased policing: because historical crime data reflects where police have focused their attention rather than where crime actually occurs, predictive models trained on that data direct police disproportionately to communities of color.
Pushback and Bans
The United States has also produced the most significant pushback against AI surveillance among major democracies. San Francisco, Boston, Minneapolis, and other cities have enacted bans or moratoriums on government use of facial recognition. Several states have passed biometric privacy laws. Federal legislation has been proposed, though not enacted, to restrict the use of facial recognition by federal agencies.
The United Kingdom: CCTV Nation
The United Kingdom has one of the highest densities of surveillance cameras in the world, with an estimated 5 to 7 million cameras across the country. The Metropolitan Police Service and other UK police forces have deployed live facial recognition technology in public spaces, scanning crowds at events, shopping districts, and transportation hubs.
The UK’s deployment of live facial recognition has been legally challenged. In 2020, the UK Court of Appeal ruled that South Wales Police’s use of facial recognition was unlawful, citing insufficient safeguards and inadequate assessment of the technology’s impact on privacy and equality. Subsequent deployments have continued under modified frameworks, but the legal and civil liberties challenges remain active.
The UK also hosts the Government Communications Headquarters (GCHQ), whose surveillance capabilities — revealed alongside NSA programs by the Snowden disclosures — include bulk interception of internet communications and AI-assisted analysis of surveillance data.
The European Union: Regulation vs. Deployment
The European Union presents a paradox: it has enacted the world’s most comprehensive AI regulation (the AI Act, which includes restrictions on biometric surveillance) while member states continue to deploy AI surveillance technology.
The AI Act prohibits real-time biometric identification in publicly accessible spaces for law enforcement purposes, with exceptions for national security threats, search for victims of crime, and prosecution of serious criminal offenses. The exceptions are broad enough to accommodate significant surveillance activity, and the implementation and enforcement of these provisions will determine whether the AI Act represents a meaningful constraint or a regulatory fig leaf.
Individual EU member states vary significantly in their surveillance practices. France has deployed facial recognition in public spaces and authorized its use for security at the 2024 Paris Olympics. Germany has been more restrictive, with strong data protection traditions and significant public opposition to biometric surveillance. Smaller EU members have adopted a range of positions.
The Middle East and North Africa
Saudi Arabia
Saudi Arabia is building comprehensive surveillance infrastructure as part of its smart city initiatives, including NEOM. The kingdom’s AI capabilities, including those being developed by HUMAIN, encompass facial recognition, social media monitoring, and data analytics platforms. Saudi Arabia’s human rights record — including the murder of journalist Jamal Khashoggi, which involved surveillance technology — raises acute concerns about the deployment of AI surveillance in a context where dissent is criminalized and independent oversight is absent.
United Arab Emirates
The UAE, particularly Dubai and Abu Dhabi, has deployed extensive AI surveillance systems, including facial recognition at airports and public spaces, smart city monitoring platforms, and social media surveillance. The UAE is also a significant exporter and investor in surveillance technology.
Israel
Israel operates sophisticated AI surveillance systems in the occupied Palestinian territories, including facial recognition at checkpoints, mobile phone surveillance, and predictive analytics. The surveillance infrastructure in the West Bank and Gaza has been described by human rights organizations as one of the most intensive in the world, applied to a population living under military occupation with no democratic recourse.
Israeli surveillance technology companies — including NSO Group, developer of the Pegasus spyware — have exported AI-powered surveillance tools to governments worldwide, including governments with documented records of using surveillance technology against journalists, activists, and political opponents.
Africa
AI surveillance technology has been deployed across the African continent, frequently through partnerships with Chinese technology companies. Huawei’s Safe City solutions have been implemented in Kenya, Uganda, Zambia, and other countries, providing integrated surveillance platforms that include facial recognition, traffic monitoring, and crime prediction.
The deployment of Chinese-supplied surveillance technology in Africa raises concerns about data sovereignty (where does the data go? who has access to it?), about the exportation of authoritarian governance models, and about the deployment of surveillance in countries with weak data protection laws and limited oversight of security services.
South Africa, Nigeria, and Kenya have emerging regulatory frameworks for data protection, but enforcement capacity is limited and the pace of surveillance deployment far exceeds the pace of regulatory development.
Southeast Asia
Singapore
Singapore operates one of the most technologically sophisticated surveillance systems in the region, with extensive CCTV networks, facial recognition deployment, and the TraceTogether contact-tracing system deployed during the COVID-19 pandemic. Singapore’s surveillance is enabled by a legal framework that provides the government broad authority and a political culture that has historically accepted higher levels of government monitoring in exchange for security and social order.
Myanmar
Myanmar’s military government has deployed Chinese-supplied facial recognition and surveillance technology to monitor opposition activists and enforce authoritarian control following the 2021 coup. Surveillance infrastructure originally deployed for civilian purposes has been repurposed for political repression.
Vietnam and Thailand
Both countries have deployed AI surveillance technology, including social media monitoring tools used to identify and suppress political dissent. Vietnam’s cybersecurity law requires technology companies to store user data domestically and provide it to the government on request, creating a legal framework for comprehensive digital surveillance.
The Supply Chain
A small number of companies supply the majority of the world’s AI surveillance technology. Understanding this supply chain is essential to understanding the global surveillance landscape.
Huawei (China) has supplied Safe City and surveillance infrastructure to at least 50 countries. Hikvision (China) is the world’s largest manufacturer of surveillance cameras. ZTE (China) has provided telecommunications surveillance infrastructure to multiple countries. NEC (Japan) supplies facial recognition and biometric systems worldwide. Palantir (US) provides data analytics platforms to intelligence agencies and law enforcement. Clearview AI (US) has supplied facial recognition to law enforcement agencies across multiple countries. NSO Group (Israel) has supplied Pegasus spyware to government clients worldwide.
The concentration of the surveillance technology supply chain in a small number of companies and countries creates leverage points for regulation. Export controls, sanctions, and procurement standards can, if implemented, constrain the global proliferation of AI surveillance technology. Whether they will be implemented depends on whether the governments that could impose such controls are willing to accept constraints on their own surveillance capabilities as the price of constraining others.
The Path Forward
The global AI surveillance infrastructure is growing faster than any effort to regulate it. The technology is becoming cheaper, more capable, and more widely available. The supply chains are globalizing. The legal frameworks that might constrain surveillance are incomplete, inconsistently enforced, and in many jurisdictions, nonexistent.
The question is not whether AI surveillance will be deployed. It is being deployed, at scale, across the world, right now. The question is whether it will be deployed with meaningful constraints — legal, institutional, and technical — that protect human rights, democratic participation, and individual privacy. Or whether the default trajectory will produce a world in which every government has the technical capacity to monitor every citizen, and the only variable is the political will to use it.
The evidence, as of early 2026, suggests that the default trajectory is winning.