Ethotechnics taxonomy

Twenty-two governance practices in six domains, each filed under the safeguards it serves, with the artifacts that apply.

Explore safeguards Safeguard matrix

Serves Authority, Correction. Applies: Standards catalog · Delegation audit · Casebook: Post Office Horizon · Theory: Democratic vs. coercive governability

Capability or practice	What it covers	Serves	Standards and tools
Capability Policy	Authoritative decision rules, constraints, and operational bounds that bind automated execution to checkable institutional commitments.	Authority, Evidence	STD-01 temporal rights Core axioms Theory: Endogenous authorization
Capability Oversight	Review bodies, escalation paths, and independent challenge circuits that keep operational power open to dispute.	Standing, Correction	STD-02 contestability and recourse Validators Casebook: Robodebt Theory: Insulation
Practice Audits	Periodic independent inspections comparing recorded decisions against evidentiary baselines and statutory grounds.	Evidence, Correction	Evidence packs Record conformance checker Theory: An engineering tradition

Serves Capability, Correction. Applies: Mechanisms catalog · Workload modeler · Theory: Ethotechnical design

Capability or practice	What it covers	Serves	Standards and tools
Capability Intake	Validation, canonical schema mapping, and verification boundaries at the point of request entry.	Capability, Evidence	STD-01 minimum binding set W3C VC schemas Theory: Friction as accidental governance
Capability Operations	Runtime orchestration, compensatory labor accounting, and exception routing across deployed automated workflows.	Correction, Dependency	Workload modeler Theory: The human subsidy to institutional continuity Casebook: England's 2020 exam grades
Practice Escalation	Automated interrupts and stop-the-line triggers that transfer execution authority to humans when operational safety thresholds trip.	Correction	Incident memo Casebook: Apple Card credit limits Theory: The sovereign override

Serves Evidence. Applies: Finite measurement tiers · The Green Dashboard working paper · Theory: What outcomes hide

Capability or practice	What it covers	Serves	Standards and tools
Capability Monitoring	Live telemetry and discrepancy tracking comparing predicted performance against realized outcomes and human rescue rates.	Evidence, Correction	Theory: Absorption as concealment Theory: The antihuman metric Casebook: Post Office Horizon
Practice Telemetry	Structured logging of decision premises, inference confidence scores, and operator workarounds for offline auditing.	Evidence	Evidence packs Theory: What outcomes hide
Capability Red teaming	Empirical failure induction and adversarial evaluation testing whether an automated system can be halted under stress.	Evidence, Capability	Delegation validity eval Frontier doctrine scan Theory: An engineering tradition
Practice Stoppability drills	Scheduled injection of synthetic edge cases and invalid authorizations to verify fail-closed behavior and alert routing.	Evidence, Capability	Stoppability evals Finite measurement tiers

Serves Standing. Applies: Start here · STD-02 contestability and recourse · Casebook: Childcare benefits scandal

Capability or practice	What it covers	Serves	Standards and tools
Capability Consent	Explicit, revocable agreements governing whether an individual's data or profile may be subjected to automated evaluation.	Standing, Dependency	Error-bearing parties Theory: Dependence runs both ways STD-01 temporal rights
Practice Consent revocation	Executable mechanisms for withdrawing consent without incurring punitive friction or losing essential public services.	Standing, Dependency	STD-01 temporal rights Theory: The model of a person is not the person
Capability Challenge routes	The operational intake channels, evidence review windows, and stop-the-line triggers through which people challenge the decisions made about them.	Standing, Correction	STD-02 contestability and recourse Theory: Challenge density Casebook: Robodebt
Practice Appeals tracks	Independent dispute tracks staffed by human decision-makers who hold the authority to stay execution and modify records.	Standing, Correction	STD-02 contestability and recourse Theory: Exception learning

Serves Authority, Capability. Applies: Authority as a lease · Laws for engineering delegated intelligence · Theory: What does not convert

Capability or practice	What it covers	Serves	Standards and tools
Capability Delegation	Explicit, time-bounded allocations of institutional decision-making authority transferred to technical delegates.	Authority	Delegation audit Justified delegation Theory: What does not convert
Practice Delegation leases	Machine-readable delegation tokens specifying non-waivable expiration dates, scope boundaries, and renewal conditions.	Authority, Capability	STD-07 Revisable delegation record Authority as a lease
Capability Policy validity	Automated attestation and runtime checking that an active delegation has not expired, exceeded scope, or lost legal backing.	Evidence, Authority	Core axioms Delegation validity eval Theory: Endogenous authorization
Capability Expansion	Treating each widening of operational scope or delegate mandate as a de novo authorization rather than a reward for past success.	Capability, Authority	Justified delegation Core axioms Theory: What does not convert

Serves Dependency. Applies: Three kinds of reversibility · Laws for engineering delegated intelligence · Theory: Automation and capture

Capability or practice	What it covers	Serves	Standards and tools
Capability Reversibility	The technical and organizational ability to unwind system state, nullify erroneous transactions, and restore prior conditions.	Correction, Dependency	Three kinds of reversibility Dependence and reversibility eval Theory: The sovereign override
Practice Rollback graphs	Automated transaction rollback graphs and restitution workflows that trigger when an underlying decision premise is refuted.	Correction, Dependency	Three kinds of reversibility STD-02 contestability and recourse
Capability Standing	Legal and procedural standing granted to captive populations who cannot exit a decision system at acceptable cost.	Standing, Dependency	Standing eval Theory: Dependence without standing Theory: Dependence runs both ways
Capability Preserved capacity	Naming, funding, and exercising the non-automated institutional capacities needed to question, replace, or halt a system.	Dependency	Error-bearing parties Laws for engineering delegated intelligence Theory: The consumption of adaptive capacity