The Ledger Comes to Life
Religions have long imagined a ledger of deeds, in which every kindness and transgression is recorded and later weighed by a divine judge. Cultural theorists Galutsky, Khutsishvili, and Aniskevich (2024) note that a modern mind might picture that ledger as an artificial intelligence keeping a “behavioral dossier” on each person. China has begun to build something close to it, and it does not wait for the afterlife.
Western debate about China’s AI-enabled surveillance usually asks what the state can see. For information professionals, a better question to ask is what the state can make people believe. A system that watches a population at scale also learns its fears and wants. In totality, such a system amounts to what I called “digital dictatorship” in research I began in 2018 (Osetrova, 2019). In an expanded version of that original publication, I developed the term into a fusion of law, data, and algorithms that governs cognition as much as conduct (Osetrova, 2025). This article argues that the same infrastructure doubles as a tool for shaping belief, first at home and increasingly abroad.
Governance Means Steering
Part of the problem is vocabulary. When Americans hear “AI governance,” they think of guardrails: oversight, audits, and accountability. The word’s older meaning is different. “Govern” descends from the Latin gubernare and the Greek kybernan, to steer a ship. Norbert Wiener (1948) drew on the same Greek root when he coined “cybernetics,” the science of control and communication.
Russian scholarship adds another layer here. Galutsky et al. (2019) observe that English spreads this territory across roughly ten words (control, management, administration, governance, operation, handling, and others), all of which can be rendered by a single Russian word, upravlenie (управление). Russian, in turn, draws a distinction that the English language blurs: upravlenie(administering a system) is not the same as rukovodstvo (руководство, leading people). Soviet thinkers took these differences seriously. In the early Soviet decades, the vocabulary of “management” gave way to rukovodstvo and titles such as “director” and “chairman,” because the older word was understood to encode a particular relationship of power. When Wiener’s Control and Communication appeared in Russian, “control” became upravlenie, and in 1969, the Soviet Academy of Sciences renamed its Institute of Automation and Telemechanics the Institute of Control Sciences (Galutsky et al., 2019). A whole society had learned that words for running things carry a theory of who runs whom.
The Chinese official language reflects the older sense. Under Xi Jinping, the Party replaced “social management” (shehui guanli) with “social governance” (shehui zhili), marking the shift of political discourse. Many Western readers perceived it as liberalization. However, Hoffman (2017) found that the two terms share the same definition and traced social management to engineering cybernetics. She describes social credit as “the attempted automation of social management.” Read this way, the system is less a score than a feedback loop: sense behavior, score it, and steer it.
This is the American blind spot. Other political cultures hear the theory inside management words; Americans tend to treat their own as technical and neutral. Galutsky et al. (2019) call this the “illusion of de-ideologization” (p. 181). “AI governance” sounds like oversight in English. Translated, packaged, and exported, the same phrase can just as easily mean steering, and U.S. policy that does not notice the difference will misread what its partners are buying.
At Home, and at Ours
The popular image of a single score for every citizen has largely proved to be a myth. MERICS researchers describe a patchwork of corporate ratings, court blacklists, and local pilots that works more like regulatory accounting than an all-seeing algorithm (Brussee, 2022). The cognitive effect, however, does not require technical completeness – it requires belief. When people assume that their purchases, activities, and contacts can be recorded and evaluated, they edit themselves in advance.
The United States is building pieces of the same architecture, usually through markets and in the name of safety rather than ideology. Until regulators stepped in, General Motors rated how its customers drove (braking, acceleration, seatbelt use) and sold the data through brokers such as LexisNexis and Verisk, which supplied insurers who could use it to set rates; many drivers did not know they had enrolled. A final Federal Trade Commission order in January 2026 now bars the practice (TechCrunch, 2026). The airlines’ own clearinghouse, the Airlines Reporting Corporation, sold Customs and Border Protection access to more than a billion passenger records under a contract that barred the agency from naming its source (Electronic Frontier Foundation, 2025). Online, 27 states require age verification to view adult content, a practice the Supreme Court upheld in Free Speech Coalition v. Paxton (2025), and six have extended similar requirements to social media (RecordingLaw, 2026). None of this amounts to a social credit system. These measures are decentralized, contested in court, and sometimes reversed, as the GM case shows. But they share its premise: everyday behavior becomes data, and data becomes a verdict about the person. Americans tend to notice that premise in Beijing and miss it at home.
Abroad: Exporting the Template
The infrastructure behind cognitive governance is now a major export. Feldstein (2019) found that Chinese firms, chiefly Huawei, Hikvision, Dahua, and ZTE, supplied AI surveillance technology to 63 countries. CSIS researchers identified 73 Huawei “Safe City” agreements across 52 countries, most with governments rated Not Free or Partly Free (Hillman & McCalpin, 2019). Doctrine travels with the hardware. Chinese officials held trainings on new media and information management for representatives of 36 of the 65 countries Freedom House surveyed, and a 2017 training for Vietnamese officials was followed by a cybersecurity law that closely mimics China’s own (Shahbaz, 2018).
The export serves Beijing’s cognitive goals in two ways. First, it normalizes the model. Each capital that adopts safe-city policing makes centralized monitoring look like ordinary modernization rather than a political choice. Second, it creates dependence. A government that relies on a foreign vendor for its cameras, servers, and software surrenders part of its control along with the contract. In effect, it rents its own information environment from someone else.
That dependence matters because the People’s Liberation Army treats the cognitive domain as a field of conflict, seeking “mind superiority” over adversary populations in peace and war (Beauchamp-Mustafaga, 2019). Such operations depend on knowing target audiences, and surveillance systems are among the richest sources of that knowledge.
Lessons for Information Professionals
- Read the vocabulary. When a partner adopts a “safe city” program, ask what is being steered and by whom. The same label can mean public safety in one capital and population control in another.
- Treat surveillance exports as influence infrastructure. Safe-city deals belong in information strategy, not only in trade or telecom security. Any partner government should be able to answer three questions about a system it buys: What does it measure? Who sees the results? And can a citizen contest what it concludes about them?
- Guard our own credibility. Just over half of advanced democracies also deploy AI surveillance (Feldstein, 2019). Democracies keep their persuasive advantage only if their own systems can be explained, audited, reviewed by a person, and shut down when they fail.
In authoritarian hands, AI does not merely watch. It steers what people believe about themselves and their state, and that capability is now for sale abroad. Cognitive warfare is no longer only what adversaries broadcast; it is also what they build.
Bio:
Inna Vladimirovna Osetrova is a graduate of the Institute of World Politics, where she earned her M.A. in Statecraft and International Affairs with a concentration in International Economics, and now leads an undergraduate academic internship program for future national security professionals. Her research focuses on AI governance and cognitive security, examining how AI and data systems shape governance and public belief. She is interested in how societies can adopt these technologies while protecting trust, transparency, and individual rights. She is a member of the Information Professionals Association and is part of the Cognitive Security Institute and The Academy.
References
Beauchamp-Mustafaga, N. (2019, September 6). Cognitive domain operations: The PLA’s new holistic concept for influence operations. China Brief, 19(16). Jamestown Foundation. https://jamestown.org/program/cognitive-domain-operations-the-plas-new-holistic-concept-for-influence-operations/
Brussee, V. (2022, February 22). China’s social credit score: Untangling myth from reality. Mercator Institute for China Studies. https://merics.org/en/opinion/chinas-social-credit-score-untangling-myth-reality
Electronic Frontier Foundation. (2025, July 9). Data brokers are selling your flight information to CBP and ICE. https://www.eff.org/deeplinks/2025/07/data-brokers-are-selling-your-flight-information-cbp-and-ice
Feldstein, S. (2019). The global expansion of AI surveillance. Carnegie Endowment for International Peace. https://carnegieendowment.org/research/2019/09/the-global-expansion-of-ai-surveillance
Free Speech Coalition, Inc. v. Paxton, 606 U.S. ___ (2025). https://www.supremecourt.gov/opinions/24pdf/23-1122_3e04.pdf
Galutsky, G. M., Fedorovich, A. V., & Khutsishvili, V. G. (2019). Stratigrafiya kul’tury i kul’turnaya politika: Osnovy teorii i metodologii [Stratigraphy of culture and cultural policy: Foundations of theory and methodology].
Galutsky, G. M., Khutsishvili, V. G., & Aniskevich, A. A. (2024). Golgofa dlya Triumfatora: Arkhitektonika, stratigrafiya, izmerenie kul’tury [Golgotha for the Triumphant: Architectonics, stratigraphy, measurement of culture]. ASKID.
Hillman, J. E., & McCalpin, M. (2019). Watching Huawei’s “safe cities”. Center for Strategic and International Studies. https://www.csis.org/analysis/watching-huaweis-safe-cities
Hoffman, S. (2017, August 17). Managing the state: Social credit, surveillance and the CCP’s plan for China. China Brief, 17(11). Jamestown Foundation. https://jamestown.org/managing-the-state-social-credit-surveillance-and-the-ccps-plan-for-china/
Osetrova, I. (2019). Digital dictatorship: The system of social credit in China. SSRN. https://doi.org/10.2139/ssrn.3828279
Osetrova, I. (2025). Digital dictatorship: Laws, algorithms, and the rise of cognitive governance. Sentinel Journal, 1(1).
RecordingLaw. (2026). Age verification laws by state (2026). https://www.recordinglaw.com/us-laws/age-verification-laws/
Shahbaz, A. (2018). Freedom on the net 2018: The rise of digital authoritarianism. Freedom House. https://freedomhouse.org/report/freedom-net/2018/rise-digital-authoritarianism
TechCrunch. (2026, January 14). The FTC’s data-sharing order against GM is finally settled. https://techcrunch.com/2026/01/14/the-ftcs-data-sharing-order-against-gm-is-finally-settled
Wiener, N. (1948). Cybernetics: Or control and communication in the animal and the machine. MIT Press.