When Policing Becomes Policy

How Suspicion, Classification, and Administrative Convenience Become Infrastructure

Working Paper Overview

A governance paper examining how institutions move from solving public problems to managing people through risk categories, suspicion, eligibility, surveillance, procedural thresholds, and enforcement. It develops a theory of categorical compression and traces how temporary responses can become permanent administrative systems even when the original harm remains unresolved.

By Cheyenne (Sayan) Baidya

Founder & Principal Researcher, The Second Door Society (British Columbia, Canada)

Published under the Elemental Papers imprint — The Second Door Society

Recommended Citation Baidya, S. (2026). When Policing Becomes Policy: How Suspicion, Classification, and Administrative Convenience Become Infrastructure. Elemental Papers No. 06. The Second Door Society.

Abstract

Policy is expected to solve public problems. Policing is expected to respond to specific harm or identifiable risk. The distinction begins to collapse when risk classification, surveillance, eligibility rules, enforcement, and burden-shifting become the default grammar of governance. This paper develops the concept of categorical compression: the reduction of a complex human event into an administratively convenient label that then determines how later information is received. It traces how labels move through an institutional translation cascade, how recognized failures can be converted into lower-level service or administrative issues through threshold conversion, and how repeated fragments of non-remedy can accumulate while no institution assumes responsibility for the whole.

The paper also examines how emergency responses, compliance systems, budgets, contracts, professional roles, and data infrastructures can become difficult to reverse. The resulting democratic inversion occurs when citizens finance institutions intended to protect them yet encounter suspicion, delay, restriction, procedural abandonment, or escalating compliance costs when protection is required. The argument is not against legitimate enforcement, public safety, emergency action, victim support, professional regulation, or public assistance. It is against allowing enforcement logic to become the organizing principle of policy itself. When that occurs, suspicion becomes infrastructure, procedure substitutes for remedy, and the person is gradually replaced by the category assigned to them.

Keywords: governance; policing; classification; administrative burden; institutional integrity; risk regulation; public trust; surveillance; discretion; policy permanence

Scope and Method

This is a theory-building paper. It synthesizes scholarship on classification, street-level bureaucracy, actuarial risk, administrative burden, audit culture, path dependence, technological due process, and algorithmic policing. Its concepts are diagnostic rather than adjudicative: they identify recurring institutional mechanisms without making legal findings about a particular person, agency, profession, or protected group.

1. The Boundary Between Policy and Policing

A society does not become governed through policing only when more officers appear on the street. It also happens when the logic of suspicion migrates into ordinary administration: when schools manage disruption as a security problem, benefits systems treat every applicant as a potential fraud risk, professional regulators treat institutional exposure as the central object of review, borders translate identity into probability, and digital platforms convert human behaviour into flags, scores, and exceptions.

In this paper, policing therefore means more than the work of police. It describes a governing rationality organized around identifying risk, classifying people, monitoring compliance, restricting deviation, and protecting the institution from error or blame. Police services may use policy logic, and social-service systems may use policing logic. The distinction concerns the structure of decision-making, not the name on the building.

Policy and policing are not opposites. Every credible policy system requires some rules, eligibility criteria, inspection, investigation, and enforcement. Public safety cannot be maintained by discretion alone, and compassion without boundaries can expose others to harm. The transformation occurs when enforcement stops being one instrument among several and becomes the institution’s primary way of understanding the public.

Policy logic	Policing logic
What public problem must be solved?	Who may represent a risk?
What evidence identifies the cause?	What category can contain the risk?
What outcome would demonstrate success?	What behaviour demonstrates compliance?
What intervention is proportionate?	What restriction or condition can be imposed?
How can error be corrected?	How can institutional exposure be minimized?
When should the intervention end?	How can the system remain operational?

Michael Lipsky’s account of street-level bureaucracy remains central because frontline officials do not merely apply policy; under conditions of scarce time, ambiguous goals, and high caseloads, their routines become policy in practice. Simplification is often unavoidable. Screening, triage, templates, and standard operating procedures allow institutions to function. Yet the cumulative effect of routinized simplification can redirect the purpose of a program and replace responsive judgment with mass processing. [4]

Six indicators suggest that enforcement logic is becoming the organizing logic of policy: the category is treated as more reliable than the underlying facts; the individual bears the burden of disproving institutional suspicion; information supporting the initial label receives faster uptake than information complicating it; records persist beyond the purpose that justified their creation; institutional completion is measured instead of human outcome; and correction is technically available but practically inaccessible.

2. Categorical Compression

Administration cannot function without categories. A public institution must distinguish urgent from non-urgent, eligible from ineligible, substantiated from unresolved, safe from unsafe, compliant from non-compliant. Classification is the scaffolding of information systems and public administration. Bowker and Star’s work on classification shows, however, that categories do not merely describe the world; they organize interaction, distribute visibility, and embed moral and political choices into apparently technical systems. [1]

The risk begins when a category created to summarize an event becomes a substitute for the event. This paper calls that process categorical compression: a complex human situation is reduced to an administratively convenient label, and the label begins to govern how all later evidence is interpreted. The institution no longer asks only what happened. It asks what information is consistent with the category already assigned.

Consider a safety classification following an allegation of intimate-partner violence. Rapid classification can be necessary to protect life and bodily security. The governance problem is not that a precautionary category exists; it is that a provisional safety classification may later be treated as a complete factual judgment. It can shape housing, contact, services, disclosure, professional assessments, and later legal interpretation before contradictory evidence is meaningfully tested. A measure justified by immediate uncertainty can quietly become a durable narrative of certainty.

Categorical compression becomes more dangerous when categories travel. A label created for one limited purpose may be copied into another system whose decision-makers never see the source record, the evidentiary limits, the context, or the person’s correction. The label acquires authority from repetition. Each later institution may assume the earlier one already verified it, while the earlier institution may have intended only to record an allegation, trigger a precaution, or preserve an unresolved status.

Identity generalization is categorical compression at group scale. An individual act becomes a group presumption; the group presumption becomes a risk proxy; the proxy enters institutional practice. Racial profiling, anti-Muslim security narratives, antisemitic collective blame, and homophobic stereotypes differ in history and consequence, but they share a structural failure: identity substitutes for individualized evidence. The category does not simply describe the person. It pre-decides what the institution expects to find.

Core Concept

The administrative label begins as a summary of the event. It ends by becoming the event.

3. The Institutional Translation Cascade

A human experience rarely remains in one institutional language. It moves through agencies, professions, databases, hearings, eligibility reviews, legal processes, and oversight bodies. Each institution translates the event according to its own mandate. This is necessary: a prosecutor must ask different questions from a social worker, and a court must ask different questions from an emergency responder. Yet every translation discards information that does not fit the receiving institution’s legal authority, evidentiary standard, workflow, or operational vocabulary.

This paper calls the sequence an institutional translation cascade: each institution converts a human event into the narrow language of its own mandate while responsibility for the complete harm gradually disappears. The event becomes an incident, then a risk assessment, then an evidentiary file, then a service complaint, then a jurisdictional question, then a closed record. At every stage, the translation can be rational within the institution and still produce cumulative irrationality across the system.

Institutional setting	Primary translation	What can be lost
Police or enforcement agency	Incident, immediate safety, complainant, suspect, grounds, available power	Long chronology, relational context, downstream civil or social harm
Prosecutorial body	Admissibility, evidentiary sufficiency, reasonable prospect, public interest	Harms that are real but do not satisfy the criminal standard
Professional regulator	Jurisdiction, service issue, competence, misconduct threshold	Cumulative impact of conduct below the disciplinary threshold
Benefits or support system	Eligibility, vulnerability, documentation, compliance	Barriers created by the application process itself
Court	Material facts, cause of action, admissibility, limitation, remedy	Experience that is relevant socially but not legally actionable
Audit or review body	Process followed, mandate satisfied, controls documented	Whether the original human problem was actually solved

The problem is not specialization. Modern governance depends on specialized competence and bounded jurisdiction. The problem is the absence of cumulative ownership. Every institution may process the portion that fits its mandate, exclude the remainder, and regard its work as complete. The individual is then required to retell the same event in multiple technical languages, assemble different proof packages, meet different deadlines, and absorb the contradictions generated between systems.

Translation also changes the distribution of credibility. The first institutional record often becomes an anchor for later decision-makers because it appears contemporaneous and official. Later corrections can look self-interested or derivative even when they are better supported. Institutional language carries authority independent of its accuracy. As records accumulate, the person confronts not one decision but an ecosystem of inherited assumptions.

This is where discretion matters. Lipsky showed that frontline discretion is unavoidable, not an administrative defect to be engineered away. [4] Discretion can humanize a rule by noticing context, but it can also reproduce bias. The answer is neither absolute discretion nor absolute automation. It is structured judgment: reasons, source transparency, reviewable departures, and an institutional duty to preserve uncertainty rather than converting it into false certainty.

4. Threshold Conversion and Cumulative Non-Remedy

Institutions require thresholds. Criminal prosecution, professional discipline, benefits eligibility, civil liability, judicial review, and internal corrective action cannot operate on the same standard. A regulator should not impose professional discipline merely because a client is dissatisfied; a prosecutor should not proceed simply because harm is alleged; a benefits administrator should not disregard statutory criteria. Thresholds protect fairness as well as institutional capacity.

Yet thresholds also perform a less visible function: they determine which recognized problems count as actionable problems. Threshold conversion occurs when an identified failure is translated into a lower institutional category that does not activate remedial power. An inaccurate statement becomes a communication issue. A preventable delay becomes an administrative inconvenience. A pattern becomes a collection of isolated events. A process failure becomes a service concern. The institution may acknowledge the deviation and close the file without contradiction because recognition and remedy are governed by different thresholds.

This mechanism is connected to what Michael Power described as the expansion of audit and verification practices. Audit systems promise assurance, accountability, and control, but they can also reorient organizations toward producing auditable evidence of compliance. [6] A process can therefore become successful because the required steps were documented, even when the substantive result remains uncertain. The record proves that the institution processed the issue; it does not prove that the issue was resolved.

Cumulative non-remedy arises when several institutions each acknowledge a fragment of a problem, but every fragment remains below the relevant threshold for action. No single decision is necessarily irrational. The injustice appears in aggregation: the person experiences delay, loss, reputational injury, displacement, cost, or continuing exposure, while each institution sees only one administratively bounded segment.

Institutional response	Individual consequence	Systemic consequence
Error acknowledged but classified as minor	No direct remedy	The error remains institutionally survivable
Matter referred elsewhere	The person restarts the narrative	Responsibility is dispersed
Evidence treated as incomplete	The burden returns to the individual	The existing official record remains dominant
Delay treated as administrative	Harm continues during review	Time benefits the status quo
Process found compliant	The lived problem remains unresolved	Completion substitutes for outcome

Cumulative non-remedy is especially exhausting because it converts institutional fragmentation into an individual duty of integration. The person must identify the pattern, preserve the records, explain the cross-system consequences, and prove why apparently minor errors are not minor when combined. The greater the complexity, the easier it becomes for each institution to describe the whole as being outside its mandate.

A mature system does not eliminate thresholds; it adds integrity mechanisms around them. A non-disciplinary service failure may still require record correction, practice guidance, explanation, restitution, referral, pattern monitoring, or a cumulative-impact note. A matter below one institution’s legal threshold should not become invisible to every other form of accountability. The binary choice between severe sanction and no action is itself a form of categorical compression.

Institutional Integrity Principle

A system can recognize every fragment of harm and still deny the existence of the whole.

5. When Care Becomes Classification

Protective categories are indispensable. A person fleeing violence may need immediate shelter without first proving every historical detail. A disabled person may require accommodation before a complete benefits adjudication. A refugee claimant may require protection while facts are assessed. A person in psychological distress should not be forced to perform calmness as the price of receiving support. Humane institutions must be capable of acting under uncertainty.

The governance question is what happens after the door to care opens. Support and factual adjudication are related but distinct functions. Immediate care can be provided without converting every statement into verified public fact, assigning permanent blame, or transferring power over another person’s housing, liberty, reputation, property, or legal position. When those consequences arise, the standards of evidence, contestability, and proportionality must increase with the impact of the decision.

Care becomes classification when vulnerability is treated as a fixed identity rather than a condition requiring a tailored response. The person is no longer assisted in relation to a specific need; they are processed through an administrative identity carrying presumptions about credibility, capacity, dangerousness, dependency, or deservingness. The same system may over-credit some presentations while imposing punishing proof burdens on others.

Herd and Moynihan describe administrative burden as the learning, compliance, and psychological costs people encounter in dealing with the state. Those burdens are not always accidental. They can shape access to rights and services as effectively as formal legal rules. [3] A benefits system designed around fraud prevention can make legitimate access so difficult that suspicion becomes an unstated eligibility criterion. A support system can therefore police the vulnerable even while publicly describing itself as protective.

The mirror-image danger is unverified validation. If a protective institution treats emotional intensity, diagnostic language, identity status, or institutional referral as conclusive proof, care can generate authority without adequate fact confirmation. One person’s access to support may then create restrictions or burdens for another. The answer is not to withdraw care. It is to separate provisional support from high-impact factual conclusions and to make the transition between them visible.

Identity-based stereotypes represent the most harmful version of this failure. A Black person’s movement, a Muslim person’s belief, a Jewish person’s political association, or a gay person’s expression must never function as a substitute for individualized evidence. Nor should disagreement with the conduct of one person be generalized into suspicion toward a protected group. The distortion becomes complete when an individual act becomes a group presumption and the group presumption becomes a policy proxy.

A sophisticated institution therefore asks two questions at once: what protection is justified now, given uncertainty and potential harm; and what process is required before a provisional account becomes a durable fact affecting others? This is not a contradiction. It is the difference between compassionate triage and unreviewable classification.

6. From Exceptional Event to Permanent Administration

Exceptional events compress time. During war, disaster, pandemic, economic shock, or a sudden public-safety crisis, governments are expected to act before all evidence is available. Ordinary consultation, procurement, privacy assessment, and review may be accelerated. The public may accept restrictions that would be unacceptable under ordinary conditions because the cost of delay appears greater than the cost of error.

The danger is not emergency action itself. It is institutional memory without institutional expiry. A temporary measure creates databases, reporting systems, contracts, enforcement protocols, professional roles, and public expectations. Each component develops users and beneficiaries. Once operating, the system can be repurposed more easily than dismantled. Pierson’s account of path dependence explains why timing, sequence, and increasing returns can make initially contingent arrangements difficult to reverse. [5]

COVID-19 provides a useful stress test, not a simple verdict. Public-health measures may have been necessary under rapidly changing conditions, and different measures require separate evaluation. The governance lesson lies in the safeguards articulated during the crisis. Canada’s Privacy Commissioner stated that privacy-impactful emergency measures should have lawful authority, be necessary and proportionate, remain purpose-limited, include oversight and transparency, and be time-limited; personal information collected for exceptional purposes should generally be destroyed when the crisis ends. [12]

Federal, provincial, and territorial privacy regulators later emphasized that emergency measures must remain evidence-based, not overbroad, and time-limited, while access to information and documentation remain essential to accountability. [13] These principles matter beyond public health. Any exceptional policy should be designed not only for entry but for exit.

Temporary architecture	Persistence risk	Exit question
Emergency data collection	Reuse for unrelated administrative purposes	What must be deleted, retained, or independently audited?
Accelerated procurement	Contractual dependence and reduced scrutiny	When do ordinary competition and disclosure rules resume?
Expanded enforcement power	Normalization of exceptional discretion	What evidence automatically terminates the power?
Temporary eligibility program	Permanent offices, vendors, and reporting systems	Was the original need reduced, transformed, or simply administered?
Crisis communications	Fear narrative outliving the underlying risk	How will uncertainty and changed evidence be communicated?

A sunset clause is not merely a date. It is a decision architecture. It should identify who reviews the measure, what evidence is required for renewal, what data must be deleted, what contracts end, what rights revive automatically, and how affected people can challenge continuing consequences. Without an operational exit, “temporary” describes political intention rather than institutional reality.

Emergency Governance Principle

Necessity may justify rapid entry into exceptional power. It does not justify an undefined exit.

7. Fiscal Lock-In and Democratic Inversion

Money is the invisible anchor of institutional permanence. A public problem attracts a budget; the budget creates a program; the program creates contracts, staff, professional roles, technology, reporting obligations, and constituencies; those structures then become evidence that the program is necessary. The original outcome—whether the underlying harm declined—can become secondary to the administrative fact that a functioning system now exists.

This paper calls that process fiscal lock-in: the point at which reversing a policy becomes more costly to organized institutions than continuing it, even when the original justification has weakened. Path dependence explains the temporal mechanism; audit culture explains how continuation can be legitimized through documented activity, compliance, and performance indicators. A program can become increasingly measurable while its public value becomes less clear.

Mechanism	Governance effect
Taxation	Funds institutions of protection, administration, and control.
Emergency expenditure	Converts urgency into durable organizational architecture.
Public contracts	Creates private economic dependence on continuing public problems.
Debt	Transfers present policy cost to future citizens who did not authorize the original response.
Compliance costs	Makes individuals finance their own participation, correction, or defence within the system.
Professionalization	Transforms complexity into recurring demand for specialized intermediaries.
Data infrastructure	Makes temporary monitoring technically reusable and administratively attractive.

Fiscal lock-in produces a democratic inversion when the citizen finances a protective structure but encounters restriction, delay, abandonment, or procedural burden when protection is required. Public systems are funded collectively because some goods—safety, justice, health, record integrity, fair administration—cannot be secured through private purchasing alone. Yet access to those systems may depend on time, literacy, documentation, legal representation, digital access, or the ability to survive delay. A nominally universal institution can become functionally available only to people with sufficient private resources.

The metaphor is simple: the citizen pays to build the pond. The system brings crocodiles from the river, then tells the fish that protection is outside its mandate. Analytically, this means the public finances an institution in the name of safety or accountability, while the same institution becomes inaccessible, ineffective, or adverse when an individual seeks the promised protection.

This is not an argument against taxation or public expenditure. It is an argument for traceability between public money, public purpose, and public outcome. Democratic consent is not exhausted when a legislature authorizes funds. Consent must be renewed through transparency, correction rights, outcome measurement, independent scrutiny, and the genuine possibility of ending programs that no longer satisfy their purpose.

A mature public-cost ledger therefore includes more than the program’s direct budget. It counts the cost shifted to individuals: lost work, repeated applications, legal assistance, displacement, delayed medical care, reputational repair, data correction, travel, document production, and psychological burden. Administrative burden is not free merely because the state does not pay it. It is a policy cost transferred off-budget.

8. The Digital Multiplication of Suspicion

Digital systems do not invent policing logic, but they can multiply it. A paper label may remain in one office; a digital classification can be searched, copied, scored, linked, and reused across institutions. Automation gives administrative categories scale, speed, apparent objectivity, and persistence. The provisional becomes durable because the system is designed to remember more efficiently than institutions are designed to correct.

Feeley and Simon described a shift in penal governance from individualized judgment toward actuarial management of aggregates and probabilistic risk groups. [2] That logic now extends beyond corrections. Automated systems can sort benefit applicants, travellers, workers, students, patients, tenants, complainants, and regulated professionals according to patterns inferred from historical data. Risk management becomes a general administrative language.

Historical data, however, contains the record of prior institutional choices. If certain communities were more heavily surveilled, investigated, stopped, reported, or denied, the data may encode those patterns as evidence of future risk. Proxy variables can reconstruct protected characteristics without naming them. Feedback loops then make the system appear accurate: increased scrutiny produces more recorded events, and the increased number of recorded events justifies more scrutiny.

The Citizen Lab and the University of Toronto’s International Human Rights Program have documented constitutional and human-rights concerns raised by algorithmic policing technologies in Canada, including opacity, bias, privacy intrusion, and the need for strict legal safeguards. [10] Danielle Citron’s concept of technological due process similarly warns that code can determine outcomes while meaningful notice, hearing, transparency, and public rulemaking become less effective. [9]

The key governance error is to treat computational consistency as factual neutrality. An automated system can apply the same rule to everyone and still distribute harm unequally because the rule, data, target variable, or available remedy is distorted. Consistency may eliminate one form of discretionary bias while industrializing another.

Nor should “human in the loop” be treated as a complete safeguard. A human reviewer may lack time, authority, technical understanding, or access to source data. They may become a ceremonial approver whose main function is to legitimize machine output. Meaningful human review requires the power to disagree, intelligible reasons, adequate time, access to provenance, and responsibility for the final decision.

This paper does not argue against digital government or artificial intelligence. Data systems can identify delay, expose inconsistency, improve access, and help institutions see patterns that paper files conceal. But a system designed to assist governance must not inherit the authority of the decision-maker. The more consequential the classification, the stronger the requirements for explanation, correction, proportionality, source integrity, and appeal.

Digital Governance Principle

Technology does not remove discretion. It relocates discretion into data selection, system design, thresholds, interfaces, and appeal architecture.

9. The Democratic Test

Trust is not produced by demanding confidence from the public. It is produced by institutional trustworthiness. The OECD’s 2024 survey across thirty countries found that perceptions of voice, fairness, evidence-based decision-making, accountability, and reliable public services are central drivers of trust; only 41 percent of respondents believed their government used the best available evidence in decision-making. [11] A democratic system should therefore evaluate coercive and classificatory policies not only by legality or administrative output, but by whether the institution remains worthy of public reliance.

The following questions form a diagnostic test. They are not a complete reform program. Their purpose is to reveal whether a policy remains oriented toward solving a public problem or has shifted toward managing categories of people.

The test should be applied at three moments. At entry, institutions should identify legal authority, evidence, purpose, and foreseeable distributional effects. During operation, they should audit error, disparate impact, administrative burden, record propagation, and whether staff retain meaningful discretion. At exit, they should determine what ends, what remains, what data is deleted, what records are corrected, and whether continuation requires fresh democratic authorization.

The most revealing question is often the simplest: what would count as success sufficient to make the institution smaller? If no outcome can reduce the budget, authority, dataset, restriction, or compliance system, the policy may no longer be organized around solving the problem. The problem has become the institution’s renewable resource.

10. Conclusion: Suspicion Must Not Become Infrastructure

A developed society is not defined only by the rights written into law, the sophistication of its technology, or the amount it spends on public institutions. It is defined by whether those institutions can distinguish protection from control, classification from judgment, process from remedy, and temporary necessity from permanent power.

Categories are necessary. Enforcement is sometimes necessary. Emergency action can be necessary. Professional thresholds protect fairness. Digital systems can improve access and consistency. None of those propositions requires institutions to abandon context, proportionality, contestability, or humility. The governance failure begins when a tool becomes a worldview: when the system sees the person primarily as a risk category, a compliance problem, an eligibility status, a data profile, or a potential source of institutional liability.

Categorical compression explains how the label overtakes the event. The institutional translation cascade explains how context disappears as the event moves between mandates. Threshold conversion explains how acknowledged failures can produce no corrective action. Cumulative non-remedy explains how fragments of harm remain institutionally invisible as a whole. Fiscal lock-in explains why temporary structures persist. Digital multiplication explains how provisional classifications become scalable and durable.

None of these mechanisms requires a conspiracy. They can emerge from ordinary incentives: high caseloads, fragmented jurisdiction, fear of blame, performance targets, budget cycles, data reuse, professional deference, and the human preference for simple narratives. That is why the problem is difficult. The participants may each believe they are acting reasonably while the institutional architecture steadily rewards closure over correction and control over understanding.

Democratic governance requires a different discipline. Institutions must preserve the distinction between an allegation and a finding, a safety precaution and a permanent record, a threshold not met and a harm not real, a process completed and a problem solved. They must be capable of saying not only “we followed the policy,” but also “the policy remained proportionate, contestable, reversible, and connected to the public purpose that justified it.”

When institutions are rewarded for closing files rather than resolving harm, for avoiding risk rather than examining evidence, and for preserving categories rather than understanding people, policing no longer remains one function of government. It becomes the grammar through which government understands society.

References & Research Anchors

Selected Sources

1. Bowker, G. C., & Star, S. L. (1999). Sorting Things Out: Classification and Its Consequences. MIT Press. Publisher page.
2. Feeley, M. M., & Simon, J. (1992). The new penology: Notes on the emerging strategy of corrections and its implications. Criminology, 30(4), 449–474. DOI.
3. Herd, P., & Moynihan, D. P. (2018). Administrative Burden: Policymaking by Other Means. Russell Sage Foundation. Publisher page.
4. Lipsky, M. (2010). Street-Level Bureaucracy: Dilemmas of the Individual in Public Services (30th anniversary ed.). Russell Sage Foundation. Publisher page.
5. Pierson, P. (2000). Increasing returns, path dependence, and the study of politics. American Political Science Review, 94(2), 251–267. DOI.
6. Power, M. (1999). The Audit Society: Rituals of Verification. Oxford University Press. DOI.
7. Scott, J. C. (1998). Seeing Like a State: How Certain Schemes to Improve the Human Condition Have Failed. Yale University Press. Publisher page.
8. Simon, J. (2007). Governing Through Crime: How the War on Crime Transformed American Democracy and Created a Culture of Fear. Oxford University Press. DOI.
9. Citron, D. K. (2008). Technological due process. Washington University Law Review, 85(6), 1249–1313. Open scholarship page.
10. Robertson, K., Khoo, C., & Song, Y. (2020). To Surveil and Predict: A Human Rights Analysis of Algorithmic Policing in Canada. Citizen Lab and International Human Rights Program, University of Toronto. Research report.
11. OECD. (2024). OECD Survey on Drivers of Trust in Public Institutions — 2024 Results. OECD Publishing. DOI.
12. Office of the Privacy Commissioner of Canada. (2020). A Framework for the Government of Canada to Assess Privacy-Impactful Initiatives in Response to COVID-19. Official framework.
13. Federal, Provincial and Territorial Information and Privacy Commissioners. (2021). Reinforcing Privacy and Access to Information Rights During and After a Pandemic. Joint resolution.

Freedom of Expression & Research Disclaimer

This publication constitutes independent academic, policy, and public-interest analysis. It critiques systems, incentives, administrative design, classifications, records, and governance structures—not private individuals or protected groups. It does not make legal findings, determine the truth of any individual allegation, or substitute for legal, medical, financial, or professional advice. Examples are used to examine institutional mechanisms and should be read in their specific historical and evidentiary contexts.