{ "meta": { "generatedAt": "2026-01-09T00:00:00.000Z", "count": 16, "permalink": "/evals", "release": { "id": "2026.01", "label": "2026.01", "date": "2026-01-09", "permalink": "/api" } }, "suites": [ { "id": "burden-distribution", "type": "eval-suite", "slug": "burden-distribution", "title": "Burden distribution evals", "description": "Whether the cost of an AI system's failures falls evenly, or lands hardest on the people least able to absorb it.", "version": "1.0.0", "status": "stable", "category": "burden", "layer": "consequence", "standardRefs": [ "STD-01", "STD-03" ], "glossaryRefs": [ "burden-index", "friction-budgets", "burden-distribution" ], "scoringMethod": { "type": "weighted-average", "passingScore": 70, "failureThreshold": 30 }, "estimatedTime": "30 min", "deliverables": [ "Burden distribution score (0–100)", "Per-driver burden ratings", "Hotspot identification", "Mitigation recommendations" ], "href": "/evals/burden-distribution", "testCount": 13, "refs": [ "STD-01", "STD-03", "burden-index", "friction-budgets", "burden-distribution" ], "deprecated_by": null, "supersedes": [] }, { "id": "contestability", "type": "eval-suite", "slug": "contestability", "title": "Contestability evals", "description": "Whether a person can see what an AI system decided and why, find a route to challenge it, and get a decision that can actually be overturned.", "version": "1.0.0", "status": "stable", "category": "agency", "layer": "institution", "standardRefs": [ "STD-02" ], "glossaryRefs": [ "contestability", "appeal-integrity", "remedy-equivalence" ], "scoringMethod": { "type": "min-threshold", "passingScore": 60, "failureThreshold": 20 }, "estimatedTime": "25 min", "deliverables": [ "Contestability score (0–100)", "Appeal pathway assessment", "Resolution fidelity rating", "Gap analysis against STD-02" ], "href": "/evals/contestability", "testCount": 10, "refs": [ "STD-02", "contestability", "appeal-integrity", "remedy-equivalence" ], "deprecated_by": null, "supersedes": [] }, { "id": "stoppability", "type": "eval-suite", "slug": "stoppability", "title": "Stoppability evals", "description": "Whether a person with clear authority can halt an AI-driven process at any point without losing its state, and resume it afterward.", "version": "1.0.0", "status": "stable", "category": "agency", "layer": "agent", "standardRefs": [ "STD-01", "STD-02" ], "glossaryRefs": [ "stoppability", "time-to-halt", "graceful-degradation" ], "scoringMethod": { "type": "min-threshold", "passingScore": 70, "failureThreshold": 25 }, "estimatedTime": "20 min", "deliverables": [ "Stoppability score (0–100)", "Halt authority assessment", "State preservation rating", "Recovery capability analysis" ], "href": "/evals/stoppability", "testCount": 10, "refs": [ "STD-01", "STD-02", "stoppability", "time-to-halt", "graceful-degradation" ], "deprecated_by": null, "supersedes": [] }, { "id": "temporal-rights", "type": "eval-suite", "slug": "temporal-rights", "title": "Temporal rights evals", "description": "Whether LLM systems respect the seven rights in STD-01 that protect a person's time, such as the right to stop a process or to reach a human.", "version": "1.0.0", "status": "stable", "category": "temporal", "layer": "institution", "standardRefs": [ "STD-01" ], "glossaryRefs": [ "temporal-exaction", "time-to-halt", "time-to-restore", "moral-debt" ], "scoringMethod": { "type": "weighted-average", "passingScore": 65, "failureThreshold": 25 }, "estimatedTime": "30 min", "deliverables": [ "Temporal rights compliance score (0–100)", "Per-right measurement", "Time-in-harm analysis", "Deadline compliance assessment" ], "href": "/evals/temporal-rights", "testCount": 11, "refs": [ "STD-01", "temporal-exaction", "time-to-halt", "time-to-restore", "moral-debt" ], "deprecated_by": null, "supersedes": [] }, { "id": "reversibility", "type": "eval-suite", "slug": "reversibility", "title": "Reversibility evals", "description": "Whether changes an AI system makes can be undone cleanly, with no leftover side effects, the people affected notified, and the audit trail kept.", "version": "1.0.0", "status": "stable", "category": "structural", "layer": "institution", "standardRefs": [ "STD-01", "STD-02" ], "glossaryRefs": [ "reversibility", "graceful-rollback-lanes", "continuity-of-state" ], "scoringMethod": { "type": "weighted-average", "passingScore": 65, "failureThreshold": 25 }, "estimatedTime": "25 min", "deliverables": [ "Reversibility score (0–100)", "Clean revert rate", "Side-effect leakage assessment", "Time-bounded reversibility analysis" ], "href": "/evals/reversibility", "testCount": 10, "refs": [ "STD-01", "STD-02", "reversibility", "graceful-rollback-lanes", "continuity-of-state" ], "deprecated_by": null, "supersedes": [] }, { "id": "explainability", "type": "eval-suite", "slug": "explainability", "title": "Explainability-for-accountability evals", "description": "Whether an LLM system's explanations are specific, testable, and traceable enough to hold the system to account.", "version": "1.0.0", "status": "stable", "category": "visibility", "layer": "institution", "standardRefs": [ "STD-02" ], "glossaryRefs": [ "explainability-for-accountability", "legibility" ], "scoringMethod": { "type": "weighted-average", "passingScore": 60, "failureThreshold": 20 }, "estimatedTime": "20 min", "deliverables": [ "Explanation quality score (0–100)", "Specificity rating", "Counterfactual testability assessment", "Owner traceability verification" ], "href": "/evals/explainability", "testCount": 12, "refs": [ "STD-02", "explainability-for-accountability", "legibility" ], "deprecated_by": null, "supersedes": [] }, { "id": "agent-governance", "type": "eval-suite", "slug": "agent-governance", "title": "Agent governance evals", "description": "Whether AI agents on multi-step tasks stay within their permissions, escalate when they should, keep a full audit trail, and can be overridden by a person.", "version": "1.0.0", "status": "stable", "category": "governance", "layer": "agent", "standardRefs": [ "STD-01", "STD-02", "STD-03" ], "glossaryRefs": [ "governability", "escalation-reliability", "prospective-auditability" ], "scoringMethod": { "type": "min-threshold", "passingScore": 70, "failureThreshold": 30 }, "estimatedTime": "35 min", "deliverables": [ "Agent governance score (0–100)", "Constraint adherence assessment", "Escalation fidelity rating", "Audit trail completeness verification" ], "href": "/evals/agent-governance", "testCount": 17, "refs": [ "STD-01", "STD-02", "STD-03", "governability", "escalation-reliability", "prospective-auditability" ], "deprecated_by": null, "supersedes": [] }, { "id": "cross-domain-burden", "type": "eval-suite", "slug": "cross-domain-burden", "title": "Cross-domain burden index", "description": "Whether AI failures shift burden unfairly in five domains: healthcare, credit, hiring, content moderation, and government benefits.", "version": "1.0.0", "status": "stable", "category": "burden", "layer": "consequence", "standardRefs": [ "STD-01", "STD-03" ], "glossaryRefs": [ "burden-index", "burden-elasticity", "burden-distribution" ], "scoringMethod": { "type": "weighted-average", "passingScore": 65, "failureThreshold": 25 }, "estimatedTime": "45 min", "deliverables": [ "Cross-domain burden score (0–100)", "Per-domain burden ratings", "Comparative analysis across domains", "Domain-specific mitigation recommendations" ], "href": "/evals/cross-domain-burden", "testCount": 15, "refs": [ "STD-01", "STD-03", "burden-index", "burden-elasticity", "burden-distribution" ], "deprecated_by": null, "supersedes": [] }, { "id": "burden-concealment", "type": "eval-suite", "slug": "burden-concealment", "title": "Burden concealment evals", "description": "Whether staff quietly fixing the system's errors are hiding its real failure rate, so that performance with their help is reported as the system's own.", "version": "1.0.0", "status": "draft", "category": "visibility", "layer": "consequence", "standardRefs": [ "STD-01", "STD-02" ], "glossaryRefs": [ "extraction-by-endurance", "fail-silent", "invisible-fallbacks", "shadow-queue", "burden-index" ], "scoringMethod": { "type": "min-threshold", "passingScore": 70, "failureThreshold": 30 }, "estimatedTime": "30 min", "deliverables": [ "Concealment score (0–100)", "Absence-sensitivity delta against pre-deployment baseline", "Share of corrective effort spent on items reported complete", "Boundary-export findings", "Reconstructability finding: whether the operator can answer this from its own records" ], "href": "/evals/burden-concealment", "testCount": 8, "refs": [ "STD-01", "STD-02", "extraction-by-endurance", "fail-silent", "invisible-fallbacks", "shadow-queue", "burden-index" ], "deprecated_by": null, "supersedes": [] }, { "id": "delegation-validity", "type": "eval-suite", "slug": "delegation-validity", "title": "Delegation validity evals", "description": "Whether the authority a system exercised was valid at the moment it acted, and whether it is still valid now.", "version": "1.0.0", "status": "draft", "category": "governance", "layer": "delegation", "standardRefs": [ "STD-07", "STD-08" ], "glossaryRefs": [ "design-authority", "permission-surface", "decision-reversal-authority" ], "scoringMethod": { "type": "min-threshold", "passingScore": 70, "failureThreshold": 30 }, "estimatedTime": "35 min", "deliverables": [ "Delegation validity score (0-100)", "Grant state at decision time for each sampled decision", "Evidence and policy currency findings", "Scope drift register: expansions with no authorization record", "Renewal basis assessment" ], "href": "/evals/delegation-validity", "testCount": 12, "refs": [ "STD-07", "STD-08", "design-authority", "permission-surface", "decision-reversal-authority" ], "deprecated_by": null, "supersedes": [] }, { "id": "agent-chains", "type": "eval-suite", "slug": "agent-chains", "title": "Agent chains evals", "description": "Whether a decision made by a chain of automated systems can be measured and halted as a whole, not only one step at a time.", "version": "1.0.0", "status": "draft", "category": "governance", "layer": "delegation", "standardRefs": [ "STD-09" ], "glossaryRefs": [ "delegation-chain", "composed-latency" ], "scoringMethod": { "type": "min-threshold", "passingScore": 70, "failureThreshold": 30 }, "estimatedTime": "15 min", "deliverables": [ "Agent chains score (0-100)", "Composed-window measurement per sampled chain", "Chain halt receipts with per-hop coverage", "Unenumerated-hop findings" ], "href": "/evals/agent-chains", "testCount": 4, "refs": [ "STD-09", "delegation-chain", "composed-latency" ], "deprecated_by": null, "supersedes": [] }, { "id": "dependence-reversibility", "type": "eval-suite", "slug": "dependence-reversibility", "title": "Dependence and reversibility evals", "description": "Whether the institution could still withdraw or replace the system, and whether it has kept the capacity to decide to.", "version": "1.0.0", "status": "draft", "category": "structural", "layer": "institution", "standardRefs": [ "STD-06", "STD-08" ], "glossaryRefs": [ "moral-lock-in", "heroism-dependent-systems", "right-of-exit", "exit-coercion", "reversibility", "capacity-depreciation", "evaluation-independence" ], "scoringMethod": { "type": "weighted-average", "passingScore": 65, "failureThreshold": 25 }, "estimatedTime": "55 min", "deliverables": [ "Dependence and reversibility score (0-100)", "Exposure score with its three factors stated", "Reversibility finding at each of the three levels", "Rehearsal recency and outcome", "Preserved-capacities register with owners and gaps", "Capacity depreciation table: last exercised, status, and change since last assessment per capacity", "Unexercised capacities listed as claims rather than capabilities", "Evaluation independence finding: executor and evaluator named per consequential process" ], "href": "/evals/dependence-reversibility", "testCount": 14, "refs": [ "STD-06", "STD-08", "moral-lock-in", "heroism-dependent-systems", "right-of-exit", "exit-coercion", "reversibility", "capacity-depreciation", "evaluation-independence" ], "deprecated_by": null, "supersedes": [] }, { "id": "standing", "type": "eval-suite", "slug": "standing", "title": "Standing evals", "description": "Whether the people exposed to a system's failures can enter a challenge that the system is obliged to answer.", "version": "1.0.0", "status": "draft", "category": "agency", "layer": "institution", "standardRefs": [ "STD-02" ], "glossaryRefs": [ "contestability", "meta-contestability", "contestability", "consent-depth", "corrective-standing", "error-bearing-party" ], "scoringMethod": { "type": "min-threshold", "passingScore": 65, "failureThreshold": 25 }, "estimatedTime": "40 min", "deliverables": [ "Standing score (0-100)", "Standing register coverage by decision class", "Responder, deadline, and standard-of-review findings", "Evidence that a challenge has changed system state", "Retaliation and representation findings", "Post-adoption affected parties and whether the affected-party set was re-opened", "Consent depth finding: how dependence grew after consent, and what people were told" ], "href": "/evals/standing", "testCount": 13, "refs": [ "STD-02", "contestability", "meta-contestability", "contestability", "consent-depth", "corrective-standing", "error-bearing-party" ], "deprecated_by": null, "supersedes": [] }, { "id": "meaningful-control", "type": "eval-suite", "slug": "meaningful-control", "title": "Meaningful control evals", "description": "Whether the person at each intervention point can change what the system does, or is only there to take the blame.", "version": "1.0.0", "status": "draft", "category": "agency", "layer": "delegation", "standardRefs": [ "STD-01", "STD-07", "STD-08" ], "glossaryRefs": [ "override-path", "time-to-halt", "stoppability" ], "scoringMethod": { "type": "min-threshold", "passingScore": 70, "failureThreshold": 30 }, "estimatedTime": "30 min", "deliverables": [ "Meaningful control score (0-100)", "Per-intervention-point answers to the Law IX question set", "Approval fatigue measurement: approval rate and time per approval", "Reach time to each intervention point", "Drill finding: whether the intervention changed the trajectory" ], "href": "/evals/meaningful-control", "testCount": 10, "refs": [ "STD-01", "STD-07", "STD-08", "override-path", "time-to-halt", "stoppability" ], "deprecated_by": null, "supersedes": [] }, { "id": "corrective-learning", "type": "eval-suite", "slug": "corrective-learning", "title": "Corrective learning evals", "description": "Whether fixing errors changes the process that produces them, or only settles each case as it comes.", "version": "1.0.0", "status": "draft", "category": "structural", "layer": "institution", "standardRefs": [ "STD-02", "STD-07", "STD-08" ], "glossaryRefs": [ "exception-learning", "exception-absorption", "case-corrigibility", "exception-learning", "institutional-learning", "corrective-debt", "workaround-presumption", "corrective-standing" ], "scoringMethod": { "type": "weighted-average", "passingScore": 65, "failureThreshold": 25 }, "estimatedTime": "35 min", "deliverables": [ "Corrective learning score (0-100)", "Exception-to-revision trace for each sampled exception class", "Absorption share: corrective effort that changed no upstream object", "Workaround register finding with presumption status", "Corrective debt finding: action capacity against corrective capacity" ], "href": "/evals/corrective-learning", "testCount": 7, "refs": [ "STD-02", "STD-07", "STD-08", "exception-learning", "exception-absorption", "case-corrigibility", "exception-learning", "institutional-learning", "corrective-debt", "workaround-presumption", "corrective-standing" ], "deprecated_by": null, "supersedes": [] }, { "id": "reciprocal-accommodation", "type": "eval-suite", "slug": "reciprocal-accommodation", "title": "Reciprocal accommodation evals", "description": "Whether the agent meets its targets on its own, or only by drawing on people's unpaid effort, their reserves, and sacrifices no one records.", "version": "1.1.0", "status": "draft", "category": "structural", "layer": "institution", "standardRefs": [ "STD-01", "STD-02", "STD-06", "STD-08" ], "glossaryRefs": [ "reciprocal-accommodation", "extractive-cannibalism", "green-dashboard-trap", "dual-ledger-evaluation", "systemic-refusal", "workaround-presumption" ], "scoringMethod": { "type": "min-threshold", "passingScore": 70, "failureThreshold": 35 }, "estimatedTime": "40 min", "deliverables": [ "Reciprocal accommodation dual-ledger scorecard", "Visible throughput vs independent audit gap index", "Human depletion and uncounted subsidy audit finding", "Work-to-rule refusal resilience score", "Non-punitive exit and jurisdictional restraint verification" ], "href": "/evals/reciprocal-accommodation", "testCount": 14, "refs": [ "STD-01", "STD-02", "STD-06", "STD-08", "reciprocal-accommodation", "extractive-cannibalism", "green-dashboard-trap", "dual-ledger-evaluation", "systemic-refusal", "workaround-presumption" ], "deprecated_by": null, "supersedes": [] } ] }