Skip to main content
⚠️ Draft — Non-authoritative content (noindex)
INFORMATIVEDRAFTnone

EU AI Act (Articles 9–15) Mapping

Normative Status

[!IMPORTANT] Informative Only This document is informative and non-normative. All mappings are illustrative only and do not imply certification, endorsement, or compliance with any external regulation.

Scope Limitation

[!WARNING] No Automatic Conformance Conformance with MPLP does not automatically guarantee conformance with the EU AI Act. Organizations must independently verify their conformance with regulatory requirements.


1. Scope

In Scope

This mapping covers EU AI Act Chapter III, Section 2 — Requirements for High-Risk AI Systems:

ArticleTitle
Article 9Risk management system
Article 10Data and data governance
Article 11Technical documentation
Article 12Record-keeping
Article 13Transparency and provision of information to deployers
Article 14Human oversight
Article 15Accuracy, robustness and cybersecurity

Out of Scope (Explicit)

  • Provider/deployer obligations beyond Articles 9–15 (e.g., quality management, post-market monitoring, incident reporting, conformity assessment)
  • Product/risk classification (Annex III applicability) and legal determinations
  • Any "certification", "badge", "endorsement", or "compliance score" framing

2. Mapping Methodology

We map each EU AI Act requirement to MPLP evidence artifacts produced by a system that follows the MPLP lifecycle:

MPLP ArtifactRole
ContextScope, environment, constraints, responsibility boundaries
PlanIntended purpose, objectives, constraints, execution phases
ConfirmHuman approvals/rejections, risk decisions, gated interventions
TraceEvent log, segment-level provenance, replayability, audit trail
RoleDeclared capabilities, permissions, operational boundaries

Interpretation Rule: MPLP does not "implement security or accuracy" by itself; it structures evidence and control points so that implementations can be audited and governed.


3. Article Mapping

Article 9 — Risk Management System

Regulatory Intent: Establish, implement, document, and maintain a continuous risk management system across the lifecycle.

MPLP Mapping: Strong

RequirementMPLP SupportMPLP Component
9.1 Establish risk management systemContext captures operational boundaries and constraintsmplp-context.schema.json
9.2(a) Identify and analyze risksPlan expresses objectives and constraintsmplp-plan.schema.json
9.2(b) Evaluate risksConfirm records risk acceptance/rejectionmplp-confirm.schema.json
9.2(c) Evaluate post-market risksTrace records monitoring signals/eventsmplp-trace.schema.json
9.2(d) Adopt risk management measuresConfirm gates enable interventionconfirm_decision_core

Evidence Type: Schema Source: schemas/v2/mplp-context.schema.json Pointer: #/properties/constraints


Article 10 — Data and Data Governance

Regulatory Intent: Ensure data quality, relevance, representativeness, and governance; address biases and gaps.

MPLP Mapping: Partial

RequirementMPLP SupportMPLP Component
10.2(a) Design choicesContext captures design backgroundcontext.root.domain
10.2(b) Data collection and originEvents record data operationscommon/events.schema.json
10.2(f) Bias detection and mitigationDrift Detection and Impact AnalysisObservability events

Note: MPLP is a protocol layer. Substantive data governance (datasets, pipelines, sampling, bias metrics) remains implementation-specific.

Evidence Type: Schema Source: schemas/v2/mplp-context.schema.json Pointer: #/properties/root


Article 11 — Technical Documentation

Regulatory Intent: Produce technical documentation before market placement, covering intended purpose, design, development, validation, and controls.

MPLP Mapping: Strong

RequirementMPLP SupportMPLP Component
Annex IV(1) Intended purposePlan objective and Context summaryplan.objective, context.summary
Annex IV(2) System architectureContext root defines domain and environmentcontext.root
Annex IV(3) Development processTrace provides complete trailmplp-trace.schema.json
Annex IV(5) Accuracy/robustness measuresConfirm decision recordsconfirm.reason
Annex IV(6) Monitoring and controlEvents and Observabilityevents[]

Evidence Type: Schema Source: schemas/v2/mplp-plan.schema.json Pointer: #/properties/objective


Article 12 — Record-keeping / Logging

Regulatory Intent: Automatic logging to enable traceability of system operations.

MPLP Mapping: Very Strong

RequirementMPLP SupportMPLP Component
12.1 Automatic event recordingTrace is designed as auditable event logmplp-trace.schema.json
12.2 Log traceabilityTrace segments support hierarchical tracingtrace_segment_coreparent_segment_id
12.3 Log retentionGovernance locked field ensures immutabilitygovernance.locked

Evidence Type: Schema Source: schemas/v2/mplp-trace.schema.json Pointer: #/properties/segments


Article 13 — Transparency and Information to Deployers

Regulatory Intent: Provide sufficient transparency for deployers to understand capabilities/limitations and use appropriately.

MPLP Mapping: Strong

RequirementMPLP SupportMPLP Component
13.1 Sufficient transparencyEvents and Trace provide explanation datacommon/events.schema.json
13.3(a) Provider identityContext owner_role defines responsibilitycontext.owner_role
13.3(b) System characteristicsRole declares capabilities and boundsmplp-role.schema.json
13.3(e) Output explanationTrace + Events reason fieldsevents[].reason

Evidence Type: Schema Source: schemas/v2/mplp-role.schema.json Pointer: #/properties (capabilities definition)


Article 14 — Human Oversight

Regulatory Intent: Effective human oversight and the ability to intervene/stop.

MPLP Mapping: Very Strong

RequirementMPLP SupportMPLP Component
14.1 Effective human oversightConfirm is the core human oversight mechanismmplp-confirm.schema.json
14.2 Intervention capabilityConfirm gates approve/reject/canceldecisions[].status
14.3(a) Understand capabilitiesRole defines capabilitiesmplp-role.schema.json
14.4(a) Identify and stop operationConfirm blocking + Plan status cancelledplan.status: cancelled

Evidence Type: Schema Source: schemas/v2/mplp-confirm.schema.json Pointer: #/$defs/confirm_decision_core


Article 15 — Accuracy, Robustness and Cybersecurity

Regulatory Intent: Achieve appropriate accuracy/robustness/cybersecurity across lifecycle; declare metrics; resist attacks.

MPLP Mapping: Partial

RequirementMPLP SupportMPLP Component
15.1 Appropriate accuracy levelsTrace provides verifiable execution recordsmplp-trace.schema.json
15.2 Resilience against errorsStatus transitions track faultsstatus: failed enums
15.3 Resist adversarial attacksRole permission constraintsmplp-role.schema.json
15.4 Feedback loop bias mitigationLearning feedback artifacts (cross-cutting duty)Learning feedback events

Note: MPLP is a protocol layer. Security implementation depends on specific runtime.

Evidence Type: Schema Source: schemas/v2/mplp-trace.schema.json Pointer: #/properties/status


4. Coverage Summary

EU AI Act ArticleMapping StrengthPrimary MPLP Artifacts
Art. 9 Risk managementStrongContext, Plan, Confirm, Trace
Art. 10 Data governancePartialTrace events, Context constraints
Art. 11 Technical documentationStrongPlan, Context, Trace, Confirm
Art. 12 Record-keepingVery StrongTrace (events + segments)
Art. 13 TransparencyStrongRole, Context/Plan, Trace
Art. 14 Human oversightVery StrongConfirm, Plan status, Trace
Art. 15 Accuracy/robustnessPartialTrace, Confirm, Role

5. Disclaimer

This mapping is provided for informational purposes only. It does not constitute legal advice or certification.

Organizations seeking EU AI Act conformity should:

  1. Consult with qualified legal and regulatory professionals
  2. Conduct independent conformity assessments
  3. Implement additional controls as required by their specific context

Related Standards: EU AI Act (Regulation (EU) 2024/1689)
See Also: ISO 42001 Mapping | NIST AI RMF Mapping