Skip to content

ML Risk and Governance

TL;DR

ML governance is the operational system that keeps model decisions accountable. It is not only compliance paperwork. It includes risk tiering, model documentation, lineage, approval gates, access control, audit logs, slice monitoring, human override, incident response, and retirement. The higher the consequence of a wrong prediction, the more the system needs explicit controls outside the model.


Risk Tiers

Not every model needs the same process.

TierExampleRequired controls
LowInternal recommendation, developer productivity rankingBasic lineage, owner, monitoring
MediumMarketing personalization, support routingExperiment review, guardrails, slice monitoring
HighFraud holds, pricing, abuse enforcementHuman override, audit logs, rollback, policy approval
CriticalCredit, employment, health, legal access decisionsFormal review, explainability, strict data governance, periodic audit

Risk tiering prevents low-risk models from being buried in bureaucracy while ensuring high-impact models do not ship as ordinary code changes.


Governance Control Plane

Governance should be encoded into the platform where possible: required metadata, promotion gates, audit logs, and model registry states.


Model Documentation

Model documentation should answer engineering and accountability questions.

SectionQuestions
Intended useWhat decision does this model support? Who uses it?
Not intended useWhere should it not be used?
Training dataWhat data, time range, labels, and exclusions were used?
FeaturesWhat feature groups and sensitive proxies are involved?
EvaluationWhich metrics, slices, and guardrails were checked?
LimitationsKnown failure cases and low-confidence domains
Human controlsReview, override, escalation, appeal
MonitoringDrift, quality, fairness/slice checks, business guardrails
OwnerTeam, on-call, review cadence

Documentation should be versioned with the model artifact, not kept in an unowned wiki page.


Data and Feature Risk

Risk often enters through data, not model code.

RiskExampleControl
Sensitive attribute useAge, health, precise locationData classification and feature allowlists
Proxy featureZIP code as proxy for protected classSlice evaluation and review
Label biasHistorical enforcement labels reflect past policyLabel audit and human review
Consent mismatchData collected for one purpose used for anotherData usage contracts
Retention violationTraining set keeps data beyond allowed windowDataset expiration and deletion workflow

Feature stores and training pipelines should enforce data classification, ownership, and retention metadata.


Human-in-the-Loop Patterns

PatternUse whenFailure mode
Human review queueModel confidence is low or action is high-impactReviewers rubber-stamp under load
Human overrideOperators need emergency controlOverrides are not logged or audited
Appeal pathUsers can contest decisionAppeal data never reaches model owners
Two-stage actionModel recommends, human decidesLatency and staffing cost
Auto-decision with auditLow-risk high-volume actionSilent bias or quality drift

Human review is a system with capacity, quality, and latency constraints. It needs metrics like any other service.


Explainability as an Operational Tool

Explainability should be tied to an action:

  • Debugging a production incident.
  • Helping reviewers make consistent decisions.
  • Supporting user-facing explanations.
  • Auditing model behavior on slices.
  • Detecting suspicious proxy features.

Explanations can be misleading if treated as truth. Use them as diagnostic artifacts alongside logs, feature values, and counterfactual tests.


Audit Log

High-impact decisions need an audit trail:

  • Request and decision timestamp.
  • Model version and policy version.
  • Input feature values or approved references.
  • Prediction, threshold, and final action.
  • Human reviewer and override reason when applicable.
  • Experiment or rollout assignment.
  • Downstream outcome and appeal result.

The audit log is also the raw material for incident analysis and retraining.


Governance Failure Modes

Unowned Model

The model runs in production but the original team moved on.

Mitigation: owner metadata, stale model alerts, review cadence, and retirement plan.

Policy Hidden in Weights

Business or safety policy is learned implicitly and cannot be reviewed.

Mitigation: keep high-impact policy constraints explicit in re-ranking, thresholds, or rules.

Metric-Only Approval

The model improves aggregate AUC but regresses a critical slice or violates a business guardrail.

Mitigation: require slice checks and guardrail approval before promotion.

No Retirement Path

Old models keep serving because nobody wants to own the migration risk.

Mitigation: model registry lifecycle states: experimental, shadow, canary, production, deprecated, retired.


Operational Metrics

CategoryMetrics
GovernanceModels by risk tier, overdue reviews, missing documentation
DataSensitive feature usage, retention violations, feature owner coverage
DecisionsOverride rate, appeal rate, review queue latency
QualitySlice metrics, calibration, false positive/negative rate
IncidentsTime to detect, time to disable model, affected decisions
LifecycleStale model age, retirement rate, rollback rate

Architecture Review Checklist

  • What risk tier is this model?
  • Who owns the model and who is on call?
  • Is model documentation versioned with the artifact?
  • Are sensitive features and proxies reviewed?
  • Are high-impact actions reversible or reviewable?
  • Is there a human override and audit path?
  • Can the model be disabled without redeploying the service?
  • Are slice metrics reviewed before and after launch?
  • Is there a retirement plan?

Key Takeaways

  1. Governance is a production control system, not a document folder.
  2. Higher-impact decisions need stronger controls outside the model.
  3. Lineage, audit logs, and ownership are reliability requirements.
  4. Explicit policy layers are easier to review than policy hidden in weights.
  5. A model without a retirement path becomes operational debt.

References

  1. Model Cards for Model Reporting
  2. Datasheets for Datasets
  3. NIST AI Risk Management Framework
  4. Hidden Technical Debt in Machine Learning Systems

A practical reference for distributed system design. Released under the MIT License.