D09 fidelity:*
Faithful disclosure / faithful representation across lifecycle
About this dimension
Server-rendered summary, fully cross-linked. The interactive workspace above lets you browse; the content here is permanent.
Compliance standards
Magnifica Humanitas (On Safeguarding the Human Person in the Time of Artificial Intelligence) · Pope Leo XIV · 2026-05-15
“fidelity to the Gospel through doctrinal development”
fidelity:epistemic_grounding
Ethics Guidelines for Trustworthy AI · High-Level Expert Group on AI (HLEG), convened by European Commission · 2019-04-08
“lifecycle responsibility; fidelity to declared purpose”
fidelity:lifecycle_application
Ethically Aligned Design, First Edition · IEEE Global Initiative on Ethics of Autonomous and Intelligent Systems · 2019-03-25
“duty-bearer obligation to fulfill rights as fidelity”
fidelity:duty_bearer_obligation_to_fulfill_rights
ASEAN Guide on AI Governance and Ethics · ASEAN Digital Ministers Meeting (ADGMIN-4) · 2024-02-02
“algorithmic disclosure; human oversight as faithful representation”
fidelity:algorithmic_disclosure + fidelity:explainability + fidelity:human_oversight_governance
CIRIS translation
Wire primitives
fidelity:*Convergence
ASEAN's fidelity-saturation is deployer-side framing (faithful disclosure to users/operators); MH's is doctrinal-epistemic. Same prefix admits both shapes.
CIRIS control
CIRIS implementation is maintained at compliance/D09_fidelity.md in the CIRISAgent repo. Open the workspace above to view the primary content inline.
Browse the full surface
Every dimension, every regulatory work and chapter, every cross-source conflict, every v1.5+ proposal — fully linked. Use this as a sitemap if you prefer to read without the interactive workspace.
27 CIRIS dimensions, by Accord principle
Beneficence (2)
Non-maleficence (2)
Integrity (13)
integrity:*
'System holds together' structural anchor — auditable, reproducible, lifecycle integrity
multilateral_participation:{forum}:{kind}
v1.3 closure for federation participation in external multilateral processes
conscience:*
Agent-side faculty layer — optimization veto, epistemic humility, coherence, alētheia
moderation:*
Federation self-correction layer (with IEEE shifting some load to partner_role:* ethics-board constructions)
method:*
Operational-design discipline (densest family overall; convergence weaker than principles — admits source-genre asymmetry honestly)
attestation:l{3,5}:*
Verification ladder (L1-L5 hardware-rooted attestation)
partner_role:*
CIRIS Registry partner-role taxonomy (ethics boards, audit bodies, stewards)
approach:*
Decision-hierarchy strategic axis (Goal→Approach→Method→Progress-Measure DAG)
progress_measure:*
Declared-metric outcomes for tracking progress toward goals
expertise:*
Declared competence in domain (named-expert attestation)
reconsideration:*
Reverse-axis appeal / rollback / negotiation-reopening primitive
key_boundary:*
CIRISEdge encryption key boundary attestation (cryptographic trust scoping)
provenance:*
Build manifest provenance (foundational technical-infrastructure attestation)
Fidelity (3)
Justice (6)
justice:*
Vulnerability-priority + fairness; tie-breaking modifier `justice:lexical_vulnerability_priority` (v1.3 CST closure) is four-source-corroborated
detection:*
LensCore F-3 / RATCHET family — aggregate-correlation / structural-injustice / drift detection
locality:decision:{scale}
v1.3 subsidiarity closure — decision routing at lowest competent scale
testimonial_witness:{kind}
v1.4 affected-party-voice closure — preserves displaced/affected/marginalized voices in attestation
witness_diversity:*
Stakeholder pluralism in design/testing/consultation
credits:*
Commons Credits substrate-building recognition (non-monetary contribution attestation)
Four regulatory works, by chapter
Magnifica Humanitas (On Safeguarding the Human Person in the Time of Artificial Intelligence)
Pope Leo XIV · 2026-05-15 · 245 atomic units mapped · 23 CIRIS dimensions attested
Chapters
- Introduction
- Ch 1 — Doctrine
- Ch 2 — Foundations
- Ch 3 — Technology & AI
- Ch 4 — Truth, Work, Freedom
- Ch 5 — Power, Love
- Conclusion
- Synthesis
Dimensions attested
- non_maleficence:*
- integrity:*
- justice:*
- prohibited:*
- detection:*
- goal:*
- locality:decision:{scale}
- autonomy:*
- fidelity:*
- beneficence:*
- multilateral_participation:{forum}:{kind}
- conscience:*
- testimonial_witness:{kind}
- witness_diversity:*
- moderation:*
- method:*
- transparency_log:*
- attestation:l{3,5}:*
- approach:*
- progress_measure:*
- expertise:*
- reconsideration:*
- credits:*
Ethics Guidelines for Trustworthy AI
High-Level Expert Group on AI (HLEG), convened by European Commission · 2019-04-08 · 133 atomic units mapped · 27 CIRIS dimensions attested
Chapters
- Introduction & Framework
- Ch I — Foundations
- Ch II — Requirements
- Ch III — Assessment
- Examples & Conclusion
Dimensions attested
- non_maleficence:*
- integrity:*
- justice:*
- prohibited:*
- detection:*
- goal:*
- locality:decision:{scale}
- autonomy:*
- fidelity:*
- beneficence:*
- multilateral_participation:{forum}:{kind}
- conscience:*
- testimonial_witness:{kind}
- witness_diversity:*
- moderation:*
- method:*
- transparency_log:*
- attestation:l{3,5}:*
- partner_role:*
- approach:*
- progress_measure:*
- expertise:*
- accountability:*
- reconsideration:*
- credits:*
- key_boundary:*
- provenance:*
Ethically Aligned Design, First Edition
IEEE Global Initiative on Ethics of Autonomous and Intelligent Systems · 2019-03-25 · 328 atomic units mapped · 27 CIRIS dimensions attested
Chapters
- Ch 01 — From Principles to Practice
- Ch 02 — General Principles
- Ch 03 — Classical Ethics
- Ch 04 — Well-Being
- Ch 05 — Affective Computing
- Ch 06 — Personal Data
- Ch 07 — Methods
- Ch 08 — Sustainable Development
- Ch 09 — Embedding Values
- Ch 10 — Policy
- Ch 11 — Law
Dimensions attested
- non_maleficence:*
- integrity:*
- justice:*
- prohibited:*
- detection:*
- goal:*
- locality:decision:{scale}
- autonomy:*
- fidelity:*
- beneficence:*
- multilateral_participation:{forum}:{kind}
- conscience:*
- testimonial_witness:{kind}
- witness_diversity:*
- moderation:*
- method:*
- transparency_log:*
- attestation:l{3,5}:*
- partner_role:*
- approach:*
- progress_measure:*
- expertise:*
- accountability:*
- reconsideration:*
- credits:*
- key_boundary:*
- provenance:*
ASEAN Guide on AI Governance and Ethics
ASEAN Digital Ministers Meeting (ADGMIN-4) · 2024-02-02 · 168 atomic units mapped · 21 CIRIS dimensions attested
Chapters
- Introduction & Principles
- Framework
- Recommendations & Conclusion
- Annex A — Risk Assessment
- Annex B — Use Cases
Dimensions attested
- non_maleficence:*
- integrity:*
- justice:*
- prohibited:*
- detection:*
- goal:*
- locality:decision:{scale}
- autonomy:*
- fidelity:*
- beneficence:*
- multilateral_participation:{forum}:{kind}
- conscience:*
- testimonial_witness:{kind}
- witness_diversity:*
- moderation:*
- method:*
- transparency_log:*
- partner_role:*
- accountability:*
- key_boundary:*
- provenance:*
Cross-source conflicts (5)
- CONF-01directHIGH
IEEE EAD Ch5 §3.4 permits licensure-gated beneficiary-deception (search/rescue, elder/child-care); CIRIS treats prohibited:deception_fraud as categorical
Disposition: Federation-level categorical posture stays. Specialization-layer consideration at CIRISMedical#1
Affects:prohibited:*
- CONF-02substrate_dispositionMEDIUM
IEEE EAD Ch5 §5 denies A/IS moral-patient candidacy; CIRIS Accord M-1 + Recursive Golden Rule + Universality Principle are substrate-agnostic
- CONF-03mutabilityMEDIUM
ASEAN §A.4.18 admits explainability fallback; other three hold explainability as constitutive at deployment time
Affects:integrity:*
- CONF-04mutabilityLOW_MEDIUM
ASEAN §A.5.3 admits opt-out where feasible; EU HLEG firmer (opt-out where possible without detriment, otherwise rectification mechanisms required)
Affects:autonomy:*
- CONF-05scope_mismatchLOW
ASEAN §A.2.1 admits experimental sandbox phases with reduced oversight; other three hold compliance constant across lifecycle stages
Affects:method:*
v1.5+ wire-format proposals (9)
Workshop proposals at CIRISRegistry#20. Not adopted.
- T3-01HIGH
detection:affective_state_shift:{axis}
- T3-02HIGH
affective:asymmetric_bond_formation:{relation}
- T3-03HIGH
nudge:delivered:{intended_behavior_axis}
- T3-04MEDIUM
detection:correlated_action:cultural_norm_drift:{population}
Preferred resolution: axis-extension on existing ecology_of_communication:* family
- T3-05MEDIUM_LOW
standing:moral_patient_candidacy:{kind}
Preferred resolution: closure-by-documentation as cross-source-conflict (see CONF-02)
- T3-06MEDIUM_HIGH
goal:planet
Affects: goal:*
- T3-07MEDIUM
partner_role:trusted_disclosure_steward:{authority}
Affects: partner_role:*
- T3-08MEDIUM
partner_role:regional_intergovernmental_working_group_dual_remit
Affects: partner_role:*
- T3-09LOW
beneficence:planetary_biosystem_intrinsic_value
Preferred resolution: extend cohort_scope enum with 'biosphere' value rather than new prefix