Educational companion dossier · Fact, interpretation, lived experience, clinical education, fiction, and mechanics are labeled separately. Scope & safety
PsychologicalWar.orgPeople before labels.
REAL-WORLD INTERPRETIVE

JUSTICE, RIGHTS & GOVERNANCE

Synthetic Cohort Simulation, Human Review, and Release Gates

An engineering and governance workflow for reproducing virtual-society stress tests, preserving failed baselines, assigning accountable reviewers, and blocking production values until both automated and human gates are satisfied.

Read the working brief Choose a learning path
LEVEL 1

ORIENTATION

Why this matters

REAL-WORLD INTERPRETIVE

One-sentence brief

A dashboard can pass while an untested cohort, inaccessible interaction, model outage, false sanction, or missing remedy still causes serious harm. Release governance must connect code, evidence, reviewer ownership, correction, rollback, and public explanation.

REAL-WORLD INTERPRETIVE

Three key points

  1. Store the model, seed, cohort definitions, thresholds, failed baselines, corrections, and digest together so a later build can reproduce the result.
  2. Automated PASS is evidence for review, not legal, clinical, accessibility, security, editorial, or deployment approval.
  3. Every material value or rule change must trigger a new run, affected human review, correction propagation, and rollback criteria.
LEVEL 2

WORKING BRIEF

Evidence, context, and limits

REAL-WORLD INTERPRETIVE

One reproducible evidence package

The release stores executable simulation code, a fixed seed, balanced cohort definitions, policy contracts, stress scenarios, acceptance checks, aggregate results, and a stable digest. A human reviewer can inspect the failed initial contract and the corrected contract instead of receiving only a final PASS label.

  • Executable method
  • Stable digest
  • Failed baseline retained
REAL-WORLD INTERPRETIVE

Thresholds are test gates, not economy settings

The model contains normalized thresholds for affordability, completion parity, dispute resolution, AI fallback, correction, appeal, repair, and reintegration. They identify unacceptable outcomes inside the test. They do not establish what a real fee, payout, grant, rent, sentence, or service-level target should be.

  • No production defaults
  • No copied report numbers
  • Candidate values require a separate decision
REAL-WORLD INTERPRETIVE

Eleven independent human gates

The matrix separates economic systems, consumer protection, accessibility, privacy, AI transparency and youth safety, clinical and lived experience, moderation and due process, regional and cultural review, security and non-actionability, editorial publication, and production deployment. Passing one does not silently approve another.

  • Named accountable role
  • Evidence-specific decision
  • Conflicts must be disclosed
REAL-WORLD INTERPRETIVE

Record approval, rejection, or insufficient evidence

A reviewer records approved with scope, rejected, changes required, or insufficient evidence. Silence, attendance, receipt of a report, automated validation, or an unsigned checklist is not approval. The repository currently records every human gate as pending and unassigned.

  • No implied consent
  • Scope belongs in the decision
  • Rejection blocks release
REAL-WORLD INTERPRETIVE

Correction and rerun are part of the design

Changes to baskets, fees, rewards, housing supply, escrow, access alternatives, AI authority, safety routing, evidence, sanctions, appeal, restitution, privacy, or public claims trigger a new digest and affected review. Historical receipts stay immutable while current records point to the superseding result.

  • Append corrections
  • Propagate to mirrors
  • Retest worst cohorts
REAL-WORLD INTERPRETIVE

Stage rollout and preserve rollback

A production candidate would require a limited rollout, observable harm indicators, stop conditions, rollback authority, incident ownership, accessible reporting, and compensation rules. The WIP.11 simulation does not authorize that rollout; it defines evidence that a future accountable owner must evaluate.

  • Stop conditions
  • Rollback owner
  • Player remedy
REAL-WORLD INTERPRETIVE

Current release state

The revised guarded contract passes all modeled scenarios and the unsafe counterfactual fails all of them. Production tuning remains blocked because human reviewers have not been assigned and real accessibility, legal, clinical, privacy, security, regional, operational, and lived-experience evidence has not been completed.

  • Automated evidence complete
  • Human approvals not granted
  • Production values unset
LEVEL 3

COMPLETE DOSSIER

Limitations, game links, and review context

DISPUTED / MULTIPLE ACCOUNTS

Known limitations and gaps

  • The simulation uses synthetic agents, normalized fixtures, and authored assumptions; it is not evidence about a real player population or a production economy.
  • No simulated fee, payout, grant, rent, wait, sentence, probability, or threshold is approved as a production value.
  • Aggregate parity can conceal smaller intersectional harms; accessibility, clinical, lived-experience, regional, legal, privacy, consumer-protection, and security review remain necessary.
  • Nationality, ethnicity, religion, language, migration, poverty, diagnosis, distress, disability identity, social identity, and real-world identity are excluded from honesty, danger, guilt, loyalty, competence, price, and enforcement decisions.
  • Automated PASS does not constitute human approval, accessibility certification, clinical validation, legal compliance, publication approval, or production readiness.
LEVEL 4

RESEARCH EDITION

Sources, methods, and stable links

REAL-WORLD INTERPRETIVE

Linked reports

  • VR MMO Economic Telemetry ResearchProtected research reference; no public report route in this release.
  • AI NPC Cost Architecture PlanProtected research reference; no public report route in this release.
REAL-WORLD VERIFIED

Method and corrections

This page follows the public method for provenance, confidence, source independence, alternative accounts, limitations, review state, and visible correction.

Read the methodView corrections
NEXT

CONTINUE

Related learning

Justice, Rights & GovernanceAfter-Action Learning, Civilian Harm, and Institutional RepairA non-humiliating after-action system that separates observation, interpretation, responsibility, uncertainty, civilian impact, repair, and learning.Justice, Rights & GovernanceIn-Game Investigations: Privacy, Evidence, and Due ProcessA server-bounded investigation model using authorized game logs, minimum necessary data, competing hypotheses, chain of custody, exculpatory evidence, independent review, and appeal.Justice, Rights & GovernanceProportionate Consequences, Due Process, and ReintegrationA conduct-based justice loop with clear rules, evidence, graduated consequences, meaningful in-custody play, appeal, restitution, sentence limits, and a real path back into ordinary participation.Justice, Rights & GovernanceSanctions, Appeals, and Virtual-Asset GovernanceA neutral moderation and asset-governance framework that uses published policy, proportionate restrictions, privacy-preserving communication, restitution, and independent appeal. Learning pathVirtual-Society Simulation and Human ReviewFollow the reproducible stress-test evidence from cross-regional cohorts through telemetry, graceful degradation, appeal, correction, reviewer ownership, and the still-closed production gate.