Executive Summary

Modern institutions increasingly rely on simulated agency: situations in which individuals are formally granted authority, choice, or responsibility, while the surrounding system pre-structures outcomes, constrains refusal, and localizes blame. This paper develops a typology of simulated agency, distinguishing it from genuine moral agency and explaining why it is both attractive to institutions and injurious to individuals.

Simulated agency is not deception in the narrow sense; it is a structural condition that emerges when legality, urgency, and fragmentation of responsibility converge. The result is a stable but morally corrosive equilibrium: systems appear accountable, individuals appear empowered, and harms remain predictable.

1. Problem Statement: Why “Agency” Requires Typology

Most ethical and organizational analyses treat agency as binary: either a person is free and responsible, or coerced and excused. This binary fails to describe contemporary institutional life, where people are:

told “it’s your call,” judged on outcomes they cannot materially control, protected by permission but exposed to moral consequence, praised for initiative and punished for restraint.

These conditions generate agency that functions symbolically rather than substantively. Without a typology, such cases are misclassified as personal failure, moral weakness, or poor judgment—masking the system dynamics that reliably produce them.

2. Defining Simulated Agency

Simulated agency exists when all of the following are present:

Formal authority is assigned to an individual role. Outcome constraints significantly limit viable choices. Refusal or delay carries asymmetric penalties. Responsibility is retrospectively individualized. Institutional structures remain insulated from consequence.

The agent appears to choose, but the system depends on that appearance to operate without redesign.

Simulated agency is thus:

ontological (a condition of being-in-a-system), structural (produced by design, not intent), recurrent (predictable across domains).

3. Why Institutions Rely on Simulated Agency

Simulated agency solves several institutional problems simultaneously:

Risk localization: harm is attached to a person, not a process. Moral laundering: legality substitutes for prudence. Decision offloading: complexity is pushed to the edge. Narrative closure: blame terminates inquiry. Operational speed: deliberation is bypassed without explicit coercion.

As a result, simulated agency is not an aberration; it is a governance technology.

4. Typology of Simulated Agency

Type I — Permission-Based Agency

“You’re allowed to do this.”

Structure: Legal or procedural permission substitutes for moral evaluation. Mechanism: Compliance is treated as ethical sufficiency. Agent experience: Relief before action; unease after outcome. Common domains: Regulatory minimums, safety floors, content moderation, medical triage thresholds.

Failure mode: The agent is blamed for outcomes that permission made thinkable but prudence would have resisted.

Type II — Urgency-Compressed Agency

“If you don’t act now, harm will occur.”

Structure: Time pressure eliminates deliberative alternatives. Mechanism: Delay is reframed as negligence. Agent experience: Moral panic, narrowed imagination, retrospective regret. Common domains: Emergency response, aviation go/no-go decisions, crisis leadership.

Failure mode: The agent is judged as if alternatives existed that urgency structurally erased.

Type III — Frontline Sovereignty

“You have final authority.”

Structure: Decision authority is placed at the lowest operational level. Mechanism: Strategic constraints remain invisible to the decision-maker. Agent experience: Isolation; responsibility without insulation. Common domains: Pilots, clinicians, moderators, pastors, site managers.

Failure mode: The individual absorbs moral injury while upstream design remains unquestioned.

Type IV — Fragmented Knowledge Agency

“You decided—based on what you knew.”

Structure: Relevant information is distributed across silos. Mechanism: No single actor sees the whole risk picture. Agent experience: Retrospective shock; sense of betrayal by the system. Common domains: Infrastructure maintenance, environmental compliance, AI deployment.

Failure mode: After harm, the agent is faulted for not integrating knowledge they were never allowed to possess.

Type V — Metric-Constrained Agency

“Meet the target; the rest is up to you.”

Structure: Quantified goals displace judgment. Mechanism: Goodhart’s Law operationalizes moral blindness. Agent experience: Strategic conformity; quiet ethical erosion. Common domains: Performance management, safety KPIs, academic administration.

Failure mode: The agent succeeds by the metric and fails by reality.

Type VI — Role-Moralized Agency

“This is what people in your role do.”

Structure: Identity-based expectations override situational prudence. Mechanism: Role loyalty becomes moral leverage. Agent experience: Guilt at refusal; exhaustion at compliance. Common domains: Clergy, caregivers, emergency professionals, public servants.

Failure mode: The agent is accused of personal failure for resisting a role script that exceeds human limits.

Type VII — Post-Hoc Responsibility Assignment

“Someone had to be responsible.”

Structure: Responsibility is assigned after outcomes are known. Mechanism: Narrative coherence replaces causal analysis. Agent experience: Confusion; moral residue without clear error. Common domains: Accident investigations, public scandals, institutional reviews.

Failure mode: The appearance of accountability prevents structural learning.

5. Distinguishing Simulated from Genuine Agency

Condition

Genuine Agency

Simulated Agency

Alternatives

Real and viable

Nominal or punitive

Refusal

Protected

Penalized

Information

Sufficient

Structurally partial

Time

Adequate

Compressed

Consequences

Proportionate

Asymmetric

Blame

Matched to control

Localized regardless

Agency becomes genuine not when choice exists in name, but when refusal is safe, information is accessible, and consequences track authority.

6. Moral Injury Without Transgression

Simulated agency explains a phenomenon otherwise misdiagnosed as burnout or weakness: moral injury incurred without wrongdoing. Individuals experience:

regret without error, guilt without culpability, exhaustion without vice.

This injury is not psychological fragility; it is ontological friction between human agency and institutional design.

7. Implications Across Domains

Aviation & safety: “Pilot error” obscures permission-based and urgency-compressed agency. Medicine: Clinician burnout reflects frontline sovereignty and metric-constrained agency. AI & technology: Moderators and engineers operate under fragmented knowledge and post-hoc blame. Church governance: Role-moralized agency substitutes calling for coercion. Environmental regulation: Permission-based compliance masks foreseeable harm.

In each case, the system functions by requiring agents to be morally exposed.

8. Why This Typology Is Resisted

Simulated agency is rarely named because it:

undermines individualist moral narratives, destabilizes blame-based accountability, implicates system designers rather than operators, demands redesign rather than training, removes the comfort of both innocence and guilt.

As Hannah Arendt observed in another context, responsibility can become everywhere acknowledged yet nowhere inhabitable. This paper argues that simulated agency is one of the primary mechanisms by which that condition is produced before catastrophe, not merely diagnosed after.

9. Conclusion: From Simulation to Substance

Simulated agency is not solved by exhortation, resilience training, or ethical reminders. It is solved only by structural changes that:

protect refusal, slow urgency, align authority with consequence, restore deliberative space, and treat restraint as competence.

Until then, institutions will continue to rely on the appearance of choice while quietly consuming the moral lives of those who appear to choose.