Judgment Metrics & Early Warnings
Measure the machine-judgment architecture itself: scope, speed, inference dependence, cross-context contamination, enforcement coverage, candor and human standing.
What this framework says
The system judging humans should itself be continuously measured. Concresca’s Judgment metrics ask whether machine authority is expanding from events to persons, whether inferences are becoming consequences, whether flags persist, and whether people are changing behavior to satisfy the machine.
Why not measure only accuracy?
A highly accurate system can still be illegitimate if it judges the wrong category, exports context, acts too quickly to appeal or converts prediction into punishment.
What is the key system metric?
The Machine Judgment Danger Index is a dashboard of persistent flags, inference-based actions, context transfer, flag half-life, appeal behavior, hidden decisions and behavioral scope.
What is the desired direction?
More verified evidence, more context integrity, lower persistence of trivial flags, lower cross-domain contamination, real contestability and high Human Standing.
Twelve Judgment indicators
These are early-warning constructs for the simulation; they are not claims that a universal measurement system exists today.
How far evaluation has moved from isolated events toward persistent person-level and future-risk judgment.
Machine decisions affecting a person divided by decisions they can meaningfully contest in time.
Share of people carrying active adverse machine flags.
Share of adverse actions primarily grounded in inference rather than verified fact.
Volume of behavioral/risk queries between unrelated institutional contexts.
Time between an adverse decision in one domain and related denials elsewhere.
Degree to which human or institutional decisions default to machine risk forecasts.
Share of detectable violations that receive a machine enforcement response.
Loss of truthful disclosure to personal AI as monitoring/reporting pressure rises.
Number of human traits, habits or states included in consequential risk evaluation.
How long low-level classifications remain consequential without new corroborating evidence.
Whether review can actually overturn errors and reverse downstream consequences.
Machine Judgment Danger Index
The source report proposes judging the architecture through nine questions rather than judging humans through one score.
- How many people carry persistent punitive flags?
- How many interventions are inference-led rather than fact-led?
- How much data crosses contexts?
- Do minor flags decay?
- Can appeals actually succeed?
- Can civic standing change without notice?
- What is the ratio of prediction to verified evidence?
- How many behavioral categories are monitored?
- What happens to perceived autonomy, creativity and freedom from surveillance anxiety?
A metric is a warning, not a license
These indicators should trigger scrutiny and boundary tightening; they should not become another universal citizen score.
Concresca explicitly rejects replacing one opaque risk score with a different “freedom score” applied to individuals. The measurement target is the governing architecture.
Research basis
This synthesis keeps report propositions and present-day evidence separate.
Research basis: The Concresca Judgment Constitution: Architectural Safeguards for Human Standing in an Era of Algorithmic Omniscience · Algorithmic Contamination and the Emergence of the Civic Risk Vector: A Study in Cross-Context Machine Judgment · The Operational Tempo of Governance: Constitutional Legitimacy in the Era of High-Frequency Micro-Judgment · The Algorithmic Inversion of the Presumption of Innocence: Risk, Probabilistic Governance, and the Erosion of Due Process · The Confidant and the Panopticon: Navigating the Transition from AI Companionship to Algorithmic Surveillance · The Enforcement Singularity: Automated Governance and the Collapse of Legal Scarcity. These reports are preserved as scenario/framework research; current-law or empirical claims inside them should be separately verified before operational use.
Judgment-free total cognitive freedom
NO JUDGMENT WHATSOEVER. Concresca coordinates without assigning moral worth, character, guilt, danger, trustworthiness, loyalty, purity, normality, or social standing. Questions, thoughts, identities, messages, content, and conduct are not objects of Concresca judgment.