VII. EVIDENCE
7.1 Evidence Rail Purpose
7.1.1 Evidence Rail as GCRI Canada’s Core Public-Good Technical Function. 7.1.1(a) The Evidence Rail shall constitute one of GCRI Canada’s core public-good technical functions and shall serve as the institutional mechanism by which GCRI Canada creates, receives, structures, reviews, compares, stewards, routes, publishes in public-safe form, corrects, supersedes, withdraws, retracts, archives, and preserves public-benefit evidence across its mandate.
7.1.1(b) The Evidence Rail shall operate within GCRI Canada’s non-executing, public-benefit, evidence, methods, observability, ontology, public-good R&D, public-good software, open technical-baseline, public authority learning, public-safe publication, validity-by-record, and correctionability role.
7.1.1(c) The Evidence Rail shall not be a recognition system, finance-readiness system, public authority system, regulatory system, procurement system, certification system, rating system, emergency command system, public warning system, investment system, insurance system, underwriting system, market system, infrastructure operation system, or execution system.
7.1.1(d) The Evidence Rail shall be designed to make evidence more trustworthy, source-lined, comparable, bounded, reviewable, challengeable, public-safe, interoperable, and correctable without converting GCRI Canada into the actor that decides, certifies, recognizes, finances, procures, regulates, commands, warns, operates, or executes.
7.1.1(e) The Evidence Rail shall support GCRI Canada’s role as an upstream technical truth and evidence integrity institution within the Nexus public-good architecture and shall remain subordinate to this Charter, applicable law, mission lock, non-execution, public-good stack alignment, public-good firewall, role separation, privacy, cybersecurity, sovereignty, public-safe publication, and correctionability.
7.1.1(f) The Evidence Rail may support internal research, public-good technical baselines, public-good software, Nexus Truth Engine methods, Nexus Observatory methods, Docket inputs, Grid inputs, Academy materials, public authority learning, GRF inputs, GRA inputs, Protocol Authority support, National Consortium interfaces, National Company interfaces, Project SPV interfaces, provider interfaces, sponsor interfaces, host interfaces, university interfaces, community interfaces, and public-safe reports only through role-bound records.
7.1.1(g) The Evidence Rail shall preserve the distinction between evidence validity and institutional consequence. Evidence may be valid, useful, decision-grade, or public-benefit-relevant without creating legal status, public authority status, recognition, finance-readiness, certification, procurement advantage, provider preference, sponsor control, public warning, emergency command, or execution consequence.
7.1.1(h) The controlling rule shall be that the Evidence Rail is GCRI Canada’s public-good technical rail for evidence integrity, not a rail for downstream authority.
7.1.2 Evidence Rail as Structured Infrastructure for Public-Benefit Evidence Across Exponential and Mission-Critical Technologies. 7.1.2(a) The Evidence Rail shall provide structured infrastructure for public-benefit evidence across all exponential, mission-critical, systemic-risk, infrastructure-relevant, resilience-relevant, public-authority-relevant, rights-affecting, public-safe, and Nexus-relevant technologies and domains within GCRI Canada’s mandate.
7.1.2(b) The Evidence Rail may apply to evidence concerning artificial intelligence, AI-RAN, O-RAN, private wireless, sovereign compute, edge compute, cloud compute, HPC, confidential computing, blockchain, DLT, Web3, DePIN, cybersecurity, cyber-physical systems, robotics, drones, sensors, Earth observation, satellite systems, geospatial intelligence, digital twins, quantum-relevant systems, biosecurity, health, climate, nature, biodiversity, WEFH systems, energy, infrastructure, semiconductors, advanced manufacturing, supply chains, space, remote connectivity, public authority systems, digital public infrastructure, financial-technology-adjacent systems, information integrity, civic infrastructure, and emerging technologies.
7.1.2(c) The Evidence Rail shall structure evidence by source, domain, evidence type, record owner where known, custodian, authority to contribute where applicable, method, date, version, geography where material, data class, confidence, uncertainty, limitations, classification, public-safe status, permitted use, prohibited use, and correction path.
7.1.2(d) The Evidence Rail shall support evidence types including observed evidence, measured evidence, sensor evidence, AI-assisted evidence, model-based evidence, simulation evidence, benchmark evidence, pilot evidence, demonstration evidence, validation sprint evidence, public authority data, community knowledge, Indigenous or local knowledge, protected knowledge, provider-supplied evidence, sponsor-supported evidence, public-good software records, technical baseline records, Observatory records, Truth Engine records, Docket records, Grid records, Rails handoff records, and publication records.
7.1.2(e) The Evidence Rail shall distinguish raw evidence, processed evidence, inferred evidence, modeled evidence, simulated evidence, synthetic evidence, human-reviewed evidence, public-safe summary, controlled annex material, restricted evidence, public-facing evidence, finance-facing input, public authority-facing material, and enterprise-stack handoff.
7.1.2(f) The Evidence Rail shall not treat all evidence as equal merely because it is recorded. Evidence shall be classified by quality, source, method, independence, confidence, uncertainty, limitations, review status, public-safe status, and correction history.
7.1.2(g) The Evidence Rail shall be capable of supporting both technology-neutral evidence discipline and technology-specific safeguard profiles where particular domains require heightened privacy, cybersecurity, public authority, finance, community, Indigenous or local knowledge, export-control, sanctions, dual-use, health, infrastructure, or public-safe treatment.
7.1.2(h) The controlling rule shall be that the Evidence Rail structures evidence across domains so that public-benefit learning can scale without losing source, context, limits, or correctionability.
7.1.3 Evidence Rail as Distinct From Recognition Rail, Finance-Readiness Rail, Protocol-Authority Rail, Public Authority Rail, Procurement Rail, and Execution Rail. 7.1.3(a) The Evidence Rail shall be institutionally, functionally, and interpretively distinct from any recognition rail, finance-readiness rail, Protocol Authority rail, public authority rail, procurement rail, and execution rail.
7.1.3(b) The Evidence Rail may provide evidence inputs to The Global Risks Forum (GRF), but such inputs shall not create GRF recognition, standing, maturity records, claims approval, registry status, public-facing legitimacy, or stakeholder formation by GCRI Canada action.
7.1.3(c) The Evidence Rail may provide evidence inputs to The Global Risks Alliance (GRA), but such inputs shall not create finance-readiness, investment advice, capital suitability, insurance-readiness, underwriting, rating, public finance approval, guarantee, capital commitment, or regulated financial consequence by GCRI Canada action.
7.1.3(d) The Evidence Rail may provide evidence, methods, ontology, technical baseline, public-good software, or proof-logic support to Nexus Standards / Protocol Authority, but such support shall not create protocol effect, conformance status, role-key authority, smart-license authority, proof-receipt authority, entitlement state, certification, or technical-validity-surface authority by GCRI Canada action.
7.1.3(e) The Evidence Rail may provide public authority learning materials or evidence handoffs to competent public authorities, but such materials shall not create public authority delegation, official guidance, regulatory finding, public service decision, public warning, emergency command, procurement approval, public finance approval, or public authority action by GCRI Canada.
7.1.3(f) The Evidence Rail may support procurement-neutral technical understanding, evidence comparability, method transparency, benchmark discipline, and public-safe claims review, but shall not create procurement specifications, tender scoring, provider prequalification, vendor ranking, contract award, public buyer recommendation, or procurement preference by default.
7.1.3(g) The Evidence Rail may support enterprise-stack actors through boundary-safe evidence handoffs, but shall not create National Company authority, Project SPV authority, provider authority, operator authority, infrastructure operation authority, market authority, regulated execution, or execution consequence by default.
7.1.3(h) The controlling rule shall be that evidence may be routed to other rails, but the authority of another rail shall not be implied from GCRI Canada evidence.
7.1.4 Evidence Rail as Support for Decision-Grade Understanding Without Becoming Decision-Making. 7.1.4(a) The Evidence Rail may support decision-grade understanding by making evidence source-lined, structured, comparable, quality-assessed, confidence-bounded, uncertainty-aware, limitation-disclosed, public-safe, reviewable, challengeable, and correctable.
7.1.4(b) “Decision-grade” shall mean structured to support decision-making by competent actors, not itself constituting a decision, approval, command, certification, recognition, finance-readiness, rating, procurement determination, public authority act, professional opinion, or execution instruction.
7.1.4(c) Decision-grade evidence artifacts may include evidence packs, technical notes, public-safe summaries, controlled annexes, benchmark records, model records, dataset records, system cards, model cards, dataset cards, benchmark cards, inference records, Docket inputs, Grid inputs, Observatory records, Truth Engine records, Academy materials, Nexus Risk Management inputs, Rails handoffs, GRF inputs, GRA inputs, and Protocol Authority support records.
7.1.4(d) Decision-grade evidence shall identify the competent actor responsible for any downstream decision where known, or shall state that downstream decisions remain with competent actors outside GCRI Canada’s default role.
7.1.4(e) The Evidence Rail shall not approve, deny, grant, refuse, rank, prioritize, allocate, inspect as authority, enforce, fine, license, permit, fund, procure, certify, warn, command, regulate, adjudicate, insure, underwrite, rate, lend, invest, guarantee, operate, or execute by virtue of making evidence decision-grade.
7.1.4(f) Where decision-grade artifacts are routed to public authorities, GRF, GRA, Protocol Authority, National Consortiums, National Companies, Project SPVs, providers, hosts, sponsors, universities, communities, or capital readers, the handoff shall include role, scope, source lineage, confidence, uncertainty, limitations, public-safe status, permitted use, prohibited use, boundary language, and correction path.
7.1.4(g) Where a decision-grade artifact is corrected, superseded, withdrawn, downgraded, reclassified, or disputed, GCRI Canada shall review downstream dependencies and notify or correct affected recipients where within its control, interface obligations, or public-safe duty.
7.1.4(h) The controlling rule shall be that the Evidence Rail may improve the quality of understanding before decisions, but it shall not make the decisions.
7.1.5 Evidence Rail as Records-Valid, Versioned, Reviewable, Challengeable, Public-Safe, and Correctionable. 7.1.5(a) The Evidence Rail shall be records-valid, versioned, reviewable, challengeable, public-safe, and correctionable. No evidence output shall be treated as institutionally valid merely because it is asserted, displayed, published, embedded in software, shown on a dashboard, referenced in a meeting, incorporated in a presentation, cited by a participant, anchored on-chain, signed, routed, or stored.
7.1.5(b) Records-valid treatment shall require source record, evidence classification, method record where material, custodian, version, date, status, confidence, uncertainty, limitations, public-safe classification, permitted use, prohibited use, and correction path.
7.1.5(c) Versioning shall apply to evidence records, evidence packs, datasets, models, system records, benchmark records, inference records, public-good software, APIs, technical baselines, dashboards, maps, reports, Docket inputs, Grid inputs, Observatory records, Truth Engine records, Rails handoffs, and public claims where material.
7.1.5(d) Reviewability shall require that material evidence may be reviewed for source integrity, method integrity, data rights, privacy, cybersecurity, sovereign data restrictions, public authority restrictions, community safeguards, protected knowledge, provider influence, sponsor influence, conflicts, public-safe status, and correction needs.
7.1.5(e) Challengeability shall require a pathway for appropriate internal or external challenge, dispute, dissent, minority view, correction request, source update, confidence change, limitation update, public-safe concern, or boundary concern.
7.1.5(f) Public-safe treatment shall require GCRI Canada to determine whether evidence may be published publicly, summarized publicly, restricted, delayed, aggregated, generalized, redacted, placed in controlled annex, routed through controlled handoff, or withheld.
7.1.5(g) Correctionability shall include correction, supersession, withdrawal, retraction, downgrade, restriction, reclassification, reinstatement, archive annotation, downstream dependency review, public-safe clarification, and closeout where appropriate.
7.1.5(h) The controlling rule shall be that evidence becomes institutionally useful only when its record, version, review status, challenge path, public-safe status, and correction path are visible.
7.1.6 Evidence Rail as Interoperable With GRF, GRA, Nexus Standards / Protocol Authority, Nexus Network, Nexus Universe, Nexus Observatory, Nexus Risk Management, Nexus Rails, Nexus Grid, Nexus Academy, Regional Nexus Consortiums, National Nexus Consortiums, National Companies, Project SPVs, Qualified Providers, Universities, Public Authorities, Hosts, Sponsors, and Communities. 7.1.6(a) The Evidence Rail shall be interoperable with relevant Nexus public-good stack and enterprise-stack interface actors, including The Global Risks Forum (GRF), The Global Risks Alliance (GRA), Nexus Standards / Protocol Authority, Nexus Network, Nexus Universe, Nexus Observatory, Nexus Risk Management, Nexus Rails, Nexus Grid, Nexus Academy, Regional Nexus Consortiums, National Nexus Consortiums, National Working Groups, National Consortium Companies, Project SPVs, qualified providers, universities, laboratories, public authorities, hosts, sponsors, donors, funders, civil society, media where appropriate, communities, and Indigenous or local knowledge holders.
7.1.6(b) Interoperability shall be achieved through records, controlled vocabulary, source-lineage discipline, evidence categories, method profiles, ontology mappings, public-safe summaries, controlled annexes, APIs where appropriate, schemas, technical baselines, public-good software, interface agreements, MoUs, handoff records, correction records, and role-bound instruments.
7.1.6(c) Interoperability shall not create legal merger, shared authority, shared treasury, shared liability, shared governance, agency, partnership, joint venture, parent-subsidiary status, public authority delegation, finance-readiness, recognition, certification, procurement preference, provider endorsement, sponsor control, protocol effect, or execution authority by default.
7.1.6(d) Evidence inputs to GRF shall preserve the distinction between GCRI Canada evidence and GRF recognition, registry, maturity, standing, claims discipline, stakeholder formation, public-facing legitimacy, and public-safe reporting functions.
7.1.6(e) Evidence inputs to GRA shall preserve the distinction between GCRI Canada evidence and GRA finance-readiness, capital readability, proof packs, insurance-readiness, capital-reader rooms, RNFD, NFD, UNFSD, and regulated-perimeter discipline.
7.1.6(f) Evidence inputs to Protocol Authority shall preserve the distinction between GCRI Canada technical evidence and protocol effect, conformance logic, role keys, smart licenses, proof receipts, entitlement states, anchoring discipline, and technical validity surfaces.
7.1.6(g) Evidence interfaces with National Companies, Project SPVs, providers, sponsors, hosts, and capital readers shall preserve public-good firewall, enterprise-stack boundary, finance-safe language, provider neutrality, sponsor non-control, competition safety, liability boundaries, public claims controls, and correction obligations.
7.1.6(h) Evidence interfaces with public authorities, universities, laboratories, civil society, media, and communities shall preserve capacity classification, non-delegation, research integrity, public-safe communication, data rights, protected knowledge, community safeguards, source protection, attribution or non-attribution, and correction paths.
7.1.6(i) The controlling rule shall be that the Evidence Rail may interoperate broadly, but interoperability shall never collapse roles.
7.1.7 Evidence Rail as Public-Good Infrastructure Protected Against Capture, Enclosure, Hype, Informal Governance, and Sponsor / Provider Influence. 7.1.7(a) The Evidence Rail shall be protected as public-good infrastructure against capture, enclosure, hype, informal governance, sponsor influence, provider influence, funder influence, host influence, public authority overreach, finance-reader pressure, media pressure, institutional drift, and private-interest substitution.
7.1.7(b) Capture shall include any attempt to shape evidence records, methods, classifications, public-safe summaries, technical baselines, dashboards, maps, publications, Docket inputs, Grid inputs, Observatory records, Truth Engine outputs, Rails handoffs, Academy materials, or public claims to favor a sponsor, provider, funder, public authority, National Company, Project SPV, capital reader, media actor, or other participant beyond the source record.
7.1.7(c) Enclosure shall include any attempt to convert public-good evidence methods, controlled vocabulary, public-good software, open technical baselines, schemas, test harnesses, public-safe methods, or correction records into proprietary control, exclusive access, locked dependency, private gatekeeping, or provider-controlled truth.
7.1.7(d) Hype shall include unsupported maturity claims, overbroad readiness claims, false certainty, exaggerated technology claims, provider-promotional framing, sponsor-promotional framing, finance-inflating language, public authority-inflating language, procurement implication, and public-facing claim inflation.
7.1.7(e) Informal governance shall include attempts to create institutional consequences through meetings, attendance, verbal statements, event participation, social media, dashboards, private messages, sponsor communications, provider claims, public authority proximity, capital-reader interest, or Nexus references without proper records and authority.
7.1.7(f) Sponsor and provider participation shall be subject to conflict controls, contribution records, influence limits, transparency where appropriate, public claims controls, provider neutrality, sponsor non-control, competition safety, public-good firewall, and correction obligations.
7.1.7(g) Where capture, enclosure, hype, informal governance, or sponsor/provider influence is detected, GCRI Canada shall hold, restrict, re-scope, correct, withdraw, reclassify, quarantine, escalate, seek review, suspend access, terminate participation, issue public-safe clarification where appropriate, and preserve records.
7.1.7(h) The controlling rule shall be that the Evidence Rail belongs to the public-good function, not to the loudest funder, provider, institution, or narrative.
7.1.8 Evidence Rail as a Mechanism for Cross-Domain Comparability Without False Equivalence. 7.1.8(a) The Evidence Rail shall serve as a mechanism for cross-domain comparability across technologies, hazards, systems, infrastructures, public authority contexts, community contexts, and Nexus interfaces without creating false equivalence.
7.1.8(b) Cross-domain comparability may include common evidence categories, confidence concepts, uncertainty concepts, source-lineage methods, maturity-context concepts, readiness-context concepts, observability categories, public-safe status labels, correction states, benchmark discipline, Docket status inputs, Grid inputs, and ontology mappings.
7.1.8(c) Comparability shall not require treating different domains, data types, hazards, technologies, communities, infrastructures, legal contexts, public authority systems, finance-facing contexts, or protected knowledge contexts as equivalent where their evidence quality, risk, meaning, authority, safeguards, or public-safe implications differ.
7.1.8(d) The Evidence Rail shall distinguish between comparable structure and comparable substance. Two records may share a template, vocabulary, confidence scale, method class, or publication format without having equal reliability, equal maturity, equal authority, equal public-safe status, equal finance relevance, or equal readiness.
7.1.8(e) Cross-domain comparisons shall identify differences in source quality, method, context, uncertainty, review status, public authority relevance, finance relevance, privacy, cybersecurity, sovereign data, community safeguards, protected knowledge, provider participation, sponsor influence, and correction history.
7.1.8(f) Comparative materials shall not imply ranking, certification, recognition, finance-readiness, procurement preference, provider endorsement, public authority approval, rating, guarantee, or market allocation unless supported by proper competent authority and record.
7.1.8(g) Where comparability creates risk of misleading public interpretation, GCRI Canada shall use narrower language, separate categories, limitation statements, public-safe design, controlled annexes, or no-public-comparison treatment.
7.1.8(h) The controlling rule shall be that the Evidence Rail may make evidence comparable enough to learn across domains, but not so flattened that differences disappear.
7.1.9 Evidence Rail as a Mechanism for Local Truth, Indigenous / Community Context, Sovereign Data, and Public Authority Context Without Erasure. 7.1.9(a) The Evidence Rail shall support local truth, Indigenous and community context, sovereign data treatment, public authority context, territorial context, cultural context, environmental context, and rights-bearing data treatment without erasure, flattening, extraction, or unsafe publication.
7.1.9(b) Local truth shall include place-specific evidence, community-specific evidence, public authority context, infrastructure context, environmental context, cultural context, language context, accessibility context, historical context, service access context, and lived or situated knowledge where appropriate and lawfully handled.
7.1.9(c) Indigenous and community context shall be handled through safeguards for consent or non-consent treatment where applicable, Indigenous data sovereignty where applicable, community protocols, protected knowledge, attribution or non-attribution, non-extraction, non-enclosure, grievance, remedy, withdrawal or restriction path, public-safe mapping, and correction path.
7.1.9(d) Sovereign data treatment shall preserve Canadian legal requirements, provincial and territorial requirements, Indigenous data sovereignty where applicable, public authority restrictions, data residency, hosting location, cross-border transfer restrictions, cloud or processor restrictions, compute-to-data needs, AI-use limits, publication limits, and correction obligations.
7.1.9(e) Public authority context shall be recorded through capacity classification, authority to contribute, official classification, permitted use, prohibited use, publication permission, reference approval, confidentiality, retention, transfer limits, correction rights, withdrawal or restriction path, and non-public-authority-action status.
7.1.9(f) The Evidence Rail shall not convert local truth, Indigenous knowledge, community knowledge, public authority data, or sovereign data into generic open data merely for interoperability, benchmarking, dashboarding, public reporting, AI training, sponsor reporting, provider testing, finance-facing materials, or public-facing visibility.
7.1.9(g) Where public-safe publication would erase context, expose sensitive places, expose protected knowledge, stigmatize communities, enable misuse, violate sovereignty, distort public authority status, or create unsafe reliance, GCRI Canada shall restrict, aggregate, generalize, redact, mask, delay, place in controlled annex, compute-to-data, localize, refuse, withdraw, or correct.
7.1.9(h) The controlling rule shall be that the Evidence Rail must make evidence interoperable without stripping the place, people, authority, sovereignty, and safeguards that give the evidence meaning.
7.1.10 Evidence Rail as a Mechanism for Public-Safe Publication and Responsible Non-Disclosure Where Required. 7.1.10(a) The Evidence Rail shall serve as a mechanism for public-safe publication and responsible non-disclosure where required by law, ethics, privacy, cybersecurity, public authority restrictions, community safeguards, Indigenous or local knowledge protections, protected knowledge limits, infrastructure sensitivity, health sensitivity, biosecurity sensitivity, national security-adjacent sensitivity, commercial sensitivity, export-control concerns, sanctions concerns, dual-use risk, public trust risk, or do-no-harm.
7.1.10(b) Public-safe publication may include public reports, public-safe summaries, technical notes, dashboards, maps, visualizations, websites, speeches, articles, social media, Academy materials, public-good software documentation, technical baseline summaries, Docket summaries, Grid summaries, Observatory summaries, Truth Engine summaries, correction notices, supersession notices, withdrawal notices, retraction notices, and public clarification notices.
7.1.10(c) Responsible non-disclosure may include withholding, controlled annex treatment, restricted access, delayed publication, aggregation, generalization, redaction, masking, de-identification where appropriate and sufficient, localization, compute-to-data, controlled rooms, clean rooms, restricted handoffs, source-protection treatment, or refusal to receive, process, retain, or publish unsafe material.
7.1.10(d) Public-safe publication shall preserve source lineage, evidence status, confidence, uncertainty, limitations, public-safe classification, non-authority language, non-warning language, non-command language, non-certification language, non-finance-readiness language, non-procurement language, non-provider-endorsement language, sponsor non-control language where material, and correction path.
7.1.10(e) Public-safe publication shall not disclose personal information, rights-bearing sensitive data, protected knowledge, community-sensitive information, public authority confidential data, health-sensitive information, biosecurity-sensitive information, cyber vulnerabilities, infrastructure weaknesses, critical dependencies, source identities, commercial secrets, controlled technology, or other sensitive material unless lawful, authorized, necessary, public-safe, and appropriately bounded.
7.1.10(f) Non-disclosure shall not be used to hide errors, avoid accountability, protect sponsor or provider interests, suppress legitimate criticism, evade correction, conceal conflicts, or prevent public-safe clarification where public trust requires disclosure within lawful and safe limits.
7.1.10(g) Where publication or non-disclosure decisions are material, the Evidence Rail shall preserve records identifying the basis for the decision, reviewer, classification, public-safe rationale, restrictions imposed, permitted use, prohibited use, affected materials, correction path, and review date.
7.1.10(h) The controlling rule shall be that public-good evidence should be made public where safe and useful, and kept controlled where disclosure would cause avoidable harm, false authority, or unsafe reliance.
7.2 Evidence as Constitutional Asset
7.2.1 Evidence as a Protected Mission Asset of GCRI Canada. 7.2.1(a) Evidence shall be a protected mission asset of GCRI Canada and shall be held, stewarded, structured, reviewed, used, published, routed, corrected, superseded, withdrawn, retracted, archived, and protected for GCRI Canada’s public-benefit purposes and not for private advantage, sponsor control, provider promotion, procurement steering, finance signaling, institutional hype, or informal authority creation.
7.2.1(b) Evidence within GCRI Canada’s custody, control, stewardship, review, publication, repository, dashboard, Docket, Grid, Observatory, Truth Engine, Rails, Academy, technical-baseline, public-good software, or interface environment shall be treated as mission-critical institutional memory and shall be subject to validity-by-record, correctionability, public-safe handling, privacy, cybersecurity, sovereign data, public authority, community safeguard, protected knowledge, IP, confidentiality, and role-separation controls.
7.2.1(c) Evidence shall support GCRI Canada’s function as an upstream technical truth, methods, observability, ontology, public-good R&D, public-good software, open technical-baseline, public authority learning, and public-safe publication institution within the Nexus public-good architecture.
7.2.1(d) Evidence shall not be treated as a disposable input, marketing asset, sponsor deliverable, provider testimonial, public authority substitute, finance-readiness signal, procurement instrument, certification artifact, public warning, emergency command, operational instruction, or execution asset.
7.2.1(e) GCRI Canada shall protect evidence from alteration, deletion, selective quotation, decontextualization, enclosure, capture, private gatekeeping, unverifiable reuse, informal governance, unsupported public claim, unsafe disclosure, and unrecorded reliance.
7.2.1(f) Evidence stewardship shall include duties of custody, classification, source preservation, context preservation, permission control, access control, versioning, reviewability, challengeability, public-safe treatment, correction, supersession, withdrawal, retraction, and archival preservation.
7.2.1(g) No person, participant, sponsor, provider, host, public authority, National Company, Project SPV, capital reader, media actor, committee, council, fellow, advisor, working group, or external actor shall use GCRI Canada evidence to create claims, authority, recognition, finance-readiness, procurement preference, certification, provider preference, sponsor control, public authority meaning, or execution consequence beyond the source record and competent authority.
7.2.1(h) The controlling rule shall be that evidence is a constitutional asset because GCRI Canada’s legitimacy depends on evidence that remains truthful, bounded, protected, and correctable.
7.2.2 Evidence as More Than Raw Data. 7.2.2(a) Evidence shall mean more than raw data. Evidence shall include data, records, observations, measurements, models, methods, context, source lineage, provenance, custody, classification, confidence, uncertainty, limitations, review status, permissions, public-safe status, correction history, and intended use.
7.2.2(b) Raw data may become evidence only when it is associated with sufficient records to identify source, collection method, data class, lawful basis or authority to contribute where applicable, custody, processing history where material, quality limits, permitted use, prohibited use, public-safe classification, and correction path.
7.2.2(c) Evidence may include sensor data, AI-RAN signals, DePIN records, geospatial layers, Earth observation imagery, cyber telemetry, dashboard records, digital twin inputs, model outputs, benchmark records, pilot records, demonstration records, validation sprint records, public authority data, community knowledge, Indigenous or local knowledge, protected knowledge, provider-supplied records, sponsor-supported records, public-good software records, technical baseline records, Docket inputs, Grid inputs, Observatory records, Truth Engine inputs, Rails handoffs, and publication records.
7.2.2(d) Evidence shall include the context necessary to understand meaning, including time, place, domain, system boundary, affected actors, source authority, method, technology used, data restrictions, community safeguards, public authority context, confidence, uncertainty, and limitations.
7.2.2(e) Evidence shall not be reduced to numbers, files, screenshots, dashboard views, model outputs, or claims without the record layer needed to interpret and correct them.
7.2.2(f) Aggregated, transformed, summarized, visualized, modeled, translated, AI-assisted, redacted, anonymized, or public-safe evidence shall retain linkage to the underlying source records or to a lawful substitute record sufficient to preserve validity and correctionability.
7.2.2(g) Where raw data lacks sufficient source, permission, context, or correction path, GCRI Canada shall classify it as incomplete, restricted, preliminary, not-for-reliance, not-for-publication, or unusable until adequate evidence records are created or obtained.
7.2.2(h) The controlling rule shall be that data becomes evidence only when it can be understood, bounded, and corrected.
7.2.3 Evidence as Record-Bearing, Source-Lined, Contextual, Classified, Permissioned, Reviewable, and Correctionable. 7.2.3(a) Evidence shall be record-bearing, source-lined, contextual, classified, permissioned, reviewable, and correctionable. No evidence shall be treated as institutionally valid merely because it is available, visible, popular, urgent, technically sophisticated, sponsor-supported, provider-supplied, public authority-adjacent, dashboarded, modeled, signed, anchored, or repeated.
7.2.3(b) Record-bearing evidence shall include a record identity, source, contributor, custodian, domain, evidence class, version, date or time period, status, access controls, public-safe status, permitted use, prohibited use, and correction path.
7.2.3(c) Source-lined evidence shall preserve the chain between the evidence and its origin, including original source, intermediate sources, transformations, summaries, translations, AI-assisted processing, model processing, dashboard processing, map processing, publication processing, and handoff processing where material.
7.2.3(d) Contextual evidence shall include sufficient domain, geographic, temporal, technical, legal, public authority, community, infrastructure, data, and safeguard context to prevent misleading use.
7.2.3(e) Classified evidence shall identify whether it is public, public-safe, internal, controlled, restricted, confidential, sensitive, health-sensitive, public authority-sensitive, infrastructure-sensitive, cyber-sensitive, community-sensitive, Indigenous or local knowledge-sensitive, protected knowledge, finance-facing, procurement-facing, or otherwise specially handled.
7.2.3(f) Permissioned evidence shall identify lawful basis or authority to receive, process, retain, publish, route, model, dashboard, map, train on, embed, summarize, translate, hand off, or archive the evidence where applicable.
7.2.3(g) Reviewable evidence shall be capable of review for source integrity, method integrity, evidence quality, privacy, cybersecurity, sovereignty, public authority boundaries, finance boundaries, community safeguards, protected knowledge, public-safe publication, claims discipline, and correction needs.
7.2.3(h) Correctionable evidence shall include a path for correction, supersession, withdrawal, retraction, downgrade, restriction, reclassification, reinstatement, archive annotation, downstream dependency review, and closeout where appropriate.
7.2.3(i) The controlling rule shall be that evidence without records, source, context, classification, permission, reviewability, and correction path is not sufficient for institutional reliance.
7.2.4 Evidence as Distinct From Assertion, Opinion, Marketing, Sponsor Narrative, Provider Claim, Public Authority Statement, Media Description, Model Output, Dashboard View, or Blockchain Entry. 7.2.4(a) Evidence shall be distinguished from assertion, opinion, marketing, sponsor narrative, provider claim, public authority statement, media description, model output, dashboard view, blockchain entry, proof receipt, social media statement, presentation, event statement, verbal assurance, meeting note, or visual indicator unless the item is supported by adequate source records, context, classification, review, and correction path.
7.2.4(b) An assertion shall not become evidence merely because it is confident, repeated, urgent, strategic, technically framed, public-facing, or made by a senior person.
7.2.4(c) An opinion shall not become evidence unless it is classified as expert opinion, community knowledge, public authority context, reviewer judgment, or other appropriate evidence class, with source, scope, basis, limitations, conflicts, and correction path recorded.
7.2.4(d) Marketing materials, sponsor narratives, provider claims, promotional statements, pitch decks, product descriptions, roadmaps, customer stories, press releases, public claims, and social media posts shall be treated as claims requiring verification, not as evidence of the truth asserted by default.
7.2.4(e) Public authority statements shall be treated according to their source, capacity, authority, date, scope, legal status, publication status, and correction path. Public authority statements shall not create GCRI Canada authority, and public authority proximity shall not transform GCRI Canada evidence into official public authority action.
7.2.4(f) Media descriptions shall be treated as public communication artifacts requiring source review, attribution review, context review, misdescription review, and correction path. Media coverage shall not create institutional validity, recognition, finance-readiness, public authority approval, certification, procurement advantage, or public trust status by default.
7.2.4(g) Model outputs, AI outputs, digital twin outputs, dashboard views, map views, sensor displays, DePIN records, blockchain anchors, hashes, smart contracts, role keys, proof receipts, and technical signals shall be treated as evidence inputs or technical records only to the extent supported by source lineage, method records, system records, confidence, uncertainty, limitations, public-safe status, and correction path.
7.2.4(h) The controlling rule shall be that evidence is not what is merely said, shown, modeled, displayed, or anchored; evidence is what is record-supported, contextualized, and correctable.
7.2.5 Evidence as Fit-for-Purpose Rather Than Absolute. 7.2.5(a) Evidence shall be evaluated as fit-for-purpose rather than absolute. GCRI Canada shall not treat evidence as universally valid, universally reliable, universally publishable, universally decision-grade, or universally reusable across domains merely because it is valid for a particular purpose.
7.2.5(b) Fit-for-purpose review shall assess the intended use, audience, domain, decision context if any, public authority context, finance-facing context, public-safe context, data class, source quality, method quality, confidence, uncertainty, limitations, legal restrictions, privacy restrictions, cybersecurity restrictions, sovereignty restrictions, community safeguards, protected knowledge restrictions, and correction path.
7.2.5(c) Evidence may be fit for research but not public publication, fit for internal learning but not public authority handoff, fit for public-safe summary but not controlled annex disclosure, fit for method development but not benchmarking, fit for benchmarking but not certification, fit for public authority learning but not public authority decision, fit for GRA input but not finance-readiness, or fit for GRF input but not recognition.
7.2.5(d) Evidence may be fit for one geography, community, technology, model, scenario, time period, dataset, provider configuration, pilot condition, demonstration condition, or institutional context and not fit for another.
7.2.5(e) Fit-for-purpose status shall be recorded where material and shall include suitable uses, unsuitable uses, public-safe limits, permitted claims, prohibited claims, and review or re-review triggers.
7.2.5(f) GCRI Canada shall not allow evidence fit for a narrow purpose to be used for broader claims, market claims, public authority claims, finance claims, procurement claims, provider claims, sponsor claims, public warning claims, certification claims, recognition claims, or execution claims.
7.2.5(g) Where evidence is reused, repurposed, transformed, routed, published, summarized, translated, modeled, embedded, dashboarded, mapped, or used in a new institutional interface, fit-for-purpose shall be reassessed.
7.2.5(h) The controlling rule shall be that evidence is valid for the purpose its record supports, not for every purpose a later actor finds convenient.
7.2.6 Evidence as Confidence-Aware and Limitation-Aware. 7.2.6(a) Evidence shall be confidence-aware and limitation-aware. GCRI Canada shall preserve confidence, uncertainty, assumptions, limitations, exclusions, data gaps, method limits, model limits, source limits, temporal limits, geographic limits, public-safe limits, and correction needs in evidence records and derivative claims.
7.2.6(b) Confidence records shall identify, where material, source quality, source independence, corroboration, method quality, review status, data completeness, measurement quality, model fit, calibration, validation, recency, dispute status, human review, public-safe status, and correction history.
7.2.6(c) Uncertainty records shall identify, where material, measurement uncertainty, model uncertainty, scenario uncertainty, classification uncertainty, source uncertainty, geospatial uncertainty, temporal uncertainty, sensor uncertainty, AI uncertainty, inference uncertainty, public authority uncertainty, community context uncertainty, and missing data.
7.2.6(d) Limitation records shall identify what the evidence does not show, what conditions were excluded, what assumptions were used, what domains are outside scope, what actors are not covered, what authority is not created, what public claims are prohibited, and what downstream uses require further review.
7.2.6(e) Evidence shall not be made to appear more certain by visual design, dashboard interface, map color, AI-generated summary, executive summary, press material, public authority proximity, sponsor framing, provider framing, or urgency language.
7.2.6(f) Where confidence is low, uncertainty is material, limitations are significant, data is sparse, evidence is disputed, or public-safe risk is high, GCRI Canada shall use restricted status, limitation language, controlled annex treatment, no-public-claim treatment, downgrade, withdrawal, or further review as appropriate.
7.2.6(g) Confidence and limitation information shall travel with evidence into evidence packs, Docket inputs, Grid inputs, Observatory records, Truth Engine records, Rails handoffs, GRF inputs, GRA inputs, Protocol Authority support, Academy materials, public authority learning materials, public-safe reports, dashboards, maps, and public claims.
7.2.6(h) The controlling rule shall be that evidence integrity requires not only stating what is known, but preserving how well it is known and where it stops.
7.2.7 Evidence as Time-Bound and Supersession-Aware. 7.2.7(a) Evidence shall be time-bound and supersession-aware. GCRI Canada shall record the date, time period, version, update status, review date, expiry or review trigger where applicable, and correction history of material evidence.
7.2.7(b) Evidence may become stale because technology changes, models change, datasets change, infrastructure changes, public authority context changes, law changes, provider status changes, sponsor context changes, community context changes, hazard conditions change, cybersecurity status changes, benchmark conditions change, pilot conditions change, or new evidence emerges.
7.2.7(c) Time-bound evidence shall not be used as if current where the source record, update frequency, review cycle, technology status, public authority context, or public-safe status indicates that re-review is required.
7.2.7(d) Supersession shall be used where a later evidence record, corrected dataset, corrected model, updated benchmark, updated technical baseline, updated dashboard, updated map, revised public-safe report, corrected public authority record, corrected community record, or corrected handoff replaces an earlier record while preserving institutional memory.
7.2.7(e) Superseded evidence shall not be deleted merely to simplify institutional memory. It shall be marked, archived, restricted where needed, and linked to the superseding record and correction path.
7.2.7(f) Where stale or superseded evidence has been used in public materials, public authority materials, finance-facing materials, procurement-facing materials, provider materials, sponsor materials, Academy materials, Docket records, Grid records, Observatory records, Truth Engine records, Rails handoffs, or public claims, GCRI Canada shall review and correct affected dependencies where material.
7.2.7(g) Evidence records shall identify review triggers, including new data, new law, new public authority position, new model version, new benchmark, new incident, new vulnerability, new community concern, new provider disclosure, new sponsor concern, new public-safe risk, or material public misunderstanding.
7.2.7(h) The controlling rule shall be that evidence lives in time, and institutional truth requires knowing when evidence was true enough, when it changed, and what replaced it.
7.2.8 Evidence as Subject to Access, Handling, Redaction, Privacy, Cybersecurity, Sovereign Data, Public Authority, and Protected Knowledge Controls. 7.2.8(a) Evidence shall be subject to access, handling, redaction, privacy, cybersecurity, sovereign data, public authority, community, Indigenous or local knowledge, protected knowledge, IP, confidentiality, infrastructure-sensitive, health-sensitive, finance-sensitive, procurement-sensitive, and public-safe controls.
7.2.8(b) Access controls shall identify who may view, process, edit, export, publish, route, summarize, model, dashboard, map, train on, embed, archive, or delete evidence, and shall be proportionate to data class, sensitivity, public authority status, public-safe status, and correction needs.
7.2.8(c) Handling controls shall identify permitted systems, repositories, data rooms, controlled rooms, clean rooms, compute environments, storage locations, processors, subprocessors, transfer paths, AI-use limits, retention, deletion or restriction obligations, audit logs, and incident procedures.
7.2.8(d) Redaction, aggregation, generalization, masking, de-identification where appropriate and sufficient, delayed release, controlled annex treatment, and responsible non-disclosure shall be used where full disclosure would create privacy harm, cybersecurity harm, public authority harm, community harm, protected knowledge exposure, infrastructure exposure, market distortion, finance misuse, procurement misuse, provider misuse, sponsor misuse, or public trust harm.
7.2.8(e) Privacy controls shall protect personal information, sensitive personal information, health-sensitive data, identity data, location data, mobility data, financial data, public service data, source identity, small-population identifiability, and inferred or derived rights-bearing information.
7.2.8(f) Cybersecurity controls shall protect evidence repositories, APIs, dashboards, maps, data rooms, clean rooms, public-good software, technical baselines, model environments, credentials, keys, secrets, access logs, vulnerability information, and incident records.
7.2.8(g) Sovereign data controls shall preserve Canadian legal requirements, provincial and territorial requirements, Indigenous data sovereignty where applicable, public authority restrictions, local community rules, data residency, hosting location, cross-border transfer limits, compute-to-data needs, AI-use limits, publication limits, and correction obligations.
7.2.8(h) Public authority controls shall preserve authority to contribute, capacity classification, official classification, permitted use, prohibited use, publication permission, reference approval, confidentiality, retention, transfer limits, correction rights, withdrawal or restriction path, and non-public-authority-action status.
7.2.8(i) Protected knowledge controls shall preserve contributor authority, community protocol, Indigenous or local knowledge status, attribution or non-attribution, permitted use, prohibited use, publication limits, public-safe mapping limits, grievance path, withdrawal or restriction path, and correction path.
7.2.8(j) The controlling rule shall be that evidence integrity includes protecting evidence from both misuse and over-disclosure.
7.2.9 Evidence as Capable of Supporting Public-Safe Knowledge Without Creating Public Authority Effect. 7.2.9(a) Evidence may support public-safe knowledge, public-benefit understanding, public authority learning, community learning, technical literacy, evidence literacy, methods literacy, risk literacy, public-good software, technical baselines, Academy materials, Docket inputs, Grid inputs, Observatory records, Truth Engine methods, Rails handoffs, GRF inputs, GRA inputs, Protocol Authority support, and public-safe reports without creating public authority effect.
7.2.9(b) Public-safe knowledge shall mean knowledge structured, bounded, reviewed, and communicated in a manner that supports public benefit while minimizing unsafe reliance, false authority, public warning confusion, emergency command confusion, finance overclaim, procurement implication, provider preference, sponsor control, privacy harm, security harm, community harm, and public trust harm.
7.2.9(c) Public-safe knowledge may be public-facing, controlled, internal, public authority-facing, community-facing, finance-facing, provider-facing, sponsor-facing, Academy-facing, Docket-facing, Grid-facing, Observatory-facing, Truth Engine-facing, or Rails-facing, provided that the proper classification, audience, limits, and correction path are recorded.
7.2.9(d) Public-safe knowledge shall not create official guidance, public authority decision, public health order, public safety directive, public warning, emergency command, regulatory finding, compliance approval, procurement approval, public finance approval, certification, recognition, finance-readiness, insurance approval, underwriting conclusion, rating, guarantee, professional advice, provider endorsement, or execution consequence by default.
7.2.9(e) Where evidence may support competent actor action, GCRI Canada may route the evidence through boundary-safe handoffs with source lineage, confidence, uncertainty, limitations, public-safe status, permitted use, prohibited use, non-authority language, and correction path.
7.2.9(f) Public authority learning materials shall preserve capacity classification, non-delegation, non-endorsement, non-official-guidance, non-warning, non-command, non-regulatory, non-procurement, non-public-finance-approval, and correction language where material.
7.2.9(g) Where public-safe knowledge is misused to imply public authority effect, GCRI Canada shall correct, restrict, withdraw, notify affected actors where appropriate, seek public authority boundary or legal review where material, and preserve correction records.
7.2.9(h) The controlling rule shall be that evidence may help the public and competent actors understand risk and systems, but understanding shall not be mistaken for public authority.
7.2.10 Evidence as Subject to Challenge, Correction, Downgrade, Withdrawal, Retraction, Retirement, and Archival. 7.2.10(a) Evidence shall remain subject to challenge, correction, downgrade, withdrawal, retraction, retirement, restriction, reclassification, reinstatement, archive annotation, downstream dependency review, and closeout.
7.2.10(b) Challenge may arise from internal review, external review, public authority concern, community concern, Indigenous or local knowledge holder concern, provider correction, sponsor correction, host correction, university correction, researcher correction, technical review, legal review, cybersecurity review, privacy review, public-safe review, finance-boundary review, procurement-boundary review, or new evidence.
7.2.10(c) Correction shall be used where evidence is inaccurate, incomplete, misleading, stale, under-contextualized, misclassified, overclaimed, wrongly attributed, insufficiently sourced, improperly permissioned, unsafe, or inconsistent with source records.
7.2.10(d) Downgrade shall be used where confidence, evidence quality, public-safe status, maturity context, readiness context, Docket status, Grid input, Observatory status, Truth Engine confidence, benchmark meaning, or public claim must be reduced because of new information, validation failure, assumption change, data gap, dispute, correction, legal concern, safeguard concern, or public-safe risk.
7.2.10(e) Withdrawal shall be used where evidence or derivative output cannot be safely corrected in place, lacks source authority, contains unauthorized data, exposes protected knowledge, reveals sensitive infrastructure, creates public authority confusion, creates public warning confusion, creates finance overclaim, creates procurement implication, creates provider preference, creates sponsor control, or creates material public trust risk.
7.2.10(f) Retraction shall be used where a public or material evidence claim was materially wrong, unsupported, unsafe, authority-inflating, finance-inflating, procurement-inflating, provider-preferential, sponsor-controlled, privacy-invasive, security-sensitive, community-harming, or inconsistent with GCRI Canada’s public-benefit duties.
7.2.10(g) Retirement shall be used where evidence is no longer active, no longer fit for present use, replaced by better evidence, outside current scope, no longer mission-relevant, or retained only for institutional memory, historical record, audit, or legal purpose.
7.2.10(h) Archival shall preserve record identity, version history, status, access restrictions, retention basis, public-safe limits, supersession status, withdrawal status, retraction status, correction history, and future retrieval limits.
7.2.10(i) Downstream dependency review shall be required where challenged, corrected, downgraded, withdrawn, retracted, retired, or archived evidence has affected public materials, public authority materials, finance-facing materials, procurement-facing materials, provider materials, sponsor materials, community-facing materials, Academy materials, Docket records, Grid records, Observatory records, Truth Engine records, Rails handoffs, GRF inputs, GRA inputs, Protocol Authority support, National Company materials, Project SPV materials, dashboards, maps, repositories, APIs, technical baselines, or public claims.
7.2.10(j) The controlling rule shall be that evidence remains trustworthy only if GCRI Canada can hear challenges, correct records, reduce confidence, withdraw unsafe claims, retract errors, retire stale records, and preserve institutional memory without pretending certainty was always present.
7.3 Evidence Categories
7.3.1 Scientific Evidence. 7.3.1(a) Scientific evidence shall include evidence derived from scientific observation, measurement, experimentation, analysis, peer-reviewed research, pre-publication research where appropriately classified, reproducible methods, laboratory work, field research, statistical analysis, environmental study, health-related study, climate study, engineering-adjacent study, computational study, and other disciplined inquiry relevant to GCRI Canada’s public-benefit mandate.
7.3.1(b) Scientific evidence shall be recorded with source, author or contributor where known, institutional context where material, method, data basis, research design, assumptions, limitations, uncertainty, confidence, peer-review or non-peer-review status, replication status where applicable, conflicts where known, public-safe status, and correction path.
7.3.1(c) Scientific evidence may support technical baselines, public-good R&D, public authority learning, Academy materials, Docket inputs, Grid inputs, Observatory methods, Truth Engine methods, GRF evidence inputs, GRA evidence inputs, Protocol Authority support, Nexus Risk Management, Nexus Rails, and public-safe reports, provided that its scope and limitations remain attached.
7.3.1(d) Scientific evidence shall not be treated as absolute truth, public authority decision, regulatory approval, certification, recognition, finance-readiness, procurement approval, provider endorsement, professional advice, public warning, emergency command, or execution instruction by default.
7.3.1(e) Scientific evidence shall distinguish peer-reviewed findings, preliminary findings, preprints, laboratory findings, field findings, observational studies, experimental studies, simulation studies, statistical associations, causal claims, expert interpretations, and public-safe summaries.
7.3.1(f) Where scientific evidence is preliminary, disputed, low-powered, non-replicated, sponsor-funded, provider-supplied, model-dependent, context-specific, stale, sensitive, or subject to material uncertainty, GCRI Canada shall classify the evidence accordingly and shall prohibit inflated claims.
7.3.1(g) Where scientific evidence is corrected, superseded, withdrawn, retracted, or materially disputed, GCRI Canada shall review dependent publications, methods, technical baselines, public-good software, Academy materials, Docket inputs, Grid inputs, Observatory records, Truth Engine records, Rails handoffs, GRF inputs, GRA inputs, public authority materials, and public claims.
7.3.1(h) The controlling rule shall be that scientific evidence strengthens GCRI Canada only when its method, uncertainty, review status, and correction history remain visible.
7.3.2 Technical Evidence. 7.3.2(a) Technical evidence shall include evidence concerning systems, architectures, software, hardware, networks, models, datasets, APIs, sensors, compute environments, cybersecurity controls, operational technology, public-good software, technical baselines, reference architectures, evaluation harnesses, test harnesses, interoperability profiles, and technical implementation contexts relevant to GCRI Canada’s mandate.
7.3.2(b) Technical evidence shall be recorded with system identity, version, configuration, owner where known, operator where applicable, custodian, provider where any, environment, dependency records, input data, output data, test conditions, benchmark conditions where applicable, security status, limitations, public-safe status, permitted use, prohibited use, and correction path.
7.3.2(c) Technical evidence may support method development, observability logic, technical baseline development, public-good software, Docket inputs, Grid inputs, Observatory records, Truth Engine methods, Protocol Authority support, public authority learning, Nexus Risk Management, Nexus Rails, and public-safe publication.
7.3.2(d) Technical evidence shall not create certification, conformance determination, engineering certification, security approval, procurement preference, provider endorsement, finance-readiness, public authority approval, operational clearance, infrastructure operation authority, or execution consequence by default.
7.3.2(e) Technical evidence shall distinguish design intent, implementation status, test result, benchmark result, production operation by another actor, pilot result, demonstration result, validation sprint result, provider assertion, reviewer finding, public-good baseline status, and public-safe summary.
7.3.2(f) Where technical evidence concerns software, APIs, public-good technical assets, repositories, or releases, the record shall identify license, maintainer, release status, dependency status, security status, known vulnerabilities, SBOM-related records where applicable, contribution rules, and correction obligations.
7.3.2(g) Where technical evidence is corrected, invalidated, superseded, withdrawn, restricted, or reclassified, GCRI Canada shall review dependent baselines, repositories, dashboards, maps, models, Academy materials, Docket records, Grid records, Observatory records, Protocol Authority support, provider materials, and public claims.
7.3.2(h) The controlling rule shall be that technical evidence must show what a system is, how it was assessed, where it works, where it fails, and what authority it does not create.
7.3.3 Observational Evidence. 7.3.3(a) Observational evidence shall include evidence derived from observation of conditions, events, systems, environments, infrastructure, communities, hazards, technologies, workflows, field contexts, public authority contexts, operational contexts by other actors, public-safe communications, and Nexus-related activities.
7.3.3(b) Observational evidence shall be recorded with observer, source, observation method, date and time period, location where material and safe, context, system or event observed, conditions, limitations, corroboration where any, confidence, uncertainty, classification, public-safe status, permitted use, prohibited use, and correction path.
7.3.3(c) Observational evidence may include field notes, structured observations, event observations, lab observations, pilot observations, demonstration observations, incident observations, public authority learning observations, community observations, host observations, provider observations, and Observatory-related observations.
7.3.3(d) Observational evidence shall not be treated as complete, representative, causal, official, certified, finance-ready, procurement-ready, provider-endorsing, public authority-approved, or execution-ready merely because it was directly observed.
7.3.3(e) Observational evidence shall distinguish firsthand observation, secondhand observation, instrument-assisted observation, AI-assisted observation, community-reported observation, provider-reported observation, public authority-reported observation, media-observed event, and public-safe summary.
7.3.3(f) Where observational evidence concerns communities, public authorities, sensitive infrastructure, protected knowledge, health-sensitive contexts, security-sensitive contexts, or public-facing claims, GCRI Canada shall apply appropriate access, privacy, public-safe, community safeguard, and correction controls.
7.3.3(g) Where observational evidence is later contradicted, corrected, contextualized, disputed, superseded, or found incomplete, GCRI Canada shall correct the record and review affected derivative materials and claims.
7.3.3(h) The controlling rule shall be that observation is evidence of what was observed under recorded conditions, not proof of all conditions not observed.
7.3.4 Sensor Evidence. 7.3.4(a) Sensor evidence shall include evidence derived from physical, environmental, infrastructure, telecommunications, health-sensitive, community, industrial, mobility, geospatial, IoT, edge, remote, wearable where lawful and appropriate, public authority, or other sensing systems within GCRI Canada’s mandate.
7.3.4(b) Sensor evidence shall be recorded with sensor identity, owner where known, operator where applicable, host, provider where any, calibration status, maintenance status where available, location or location-treatment, timing, data format, collection method, data quality, tamper risk, spoof risk, uncertainty, limitations, classification, public-safe status, permitted use, prohibited use, and correction path.
7.3.4(c) Sensor evidence may support observability methods, Nexus Observatory records, AI-RAN signal evidence, DePIN evidence, environmental evidence, climate evidence, infrastructure evidence, WEFH evidence, public authority learning, digital twin inputs, Truth Engine comparison, Docket inputs, Grid inputs, and public-safe dashboards or maps.
7.3.4(d) Sensor evidence shall not create official measurement status, public authority finding, public warning, emergency command, certification, provider endorsement, finance-readiness, procurement approval, infrastructure operation authority, or execution consequence by default.
7.3.4(e) Sensor evidence shall distinguish raw sensor reading, calibrated reading, processed signal, fused signal, inferred condition, modeled condition, public-safe visualization, dashboard output, map layer, and handoff record.
7.3.4(f) Where sensor evidence involves location sensitivity, health sensitivity, public authority data, community data, Indigenous or local knowledge, infrastructure sensitivity, cybersecurity risk, or small-population identifiability, heightened privacy, public-safe, sovereignty, and safeguard controls shall apply.
7.3.4(g) Where a sensor is recalibrated, found faulty, spoofed, tampered with, mislocated, misconfigured, stale, replaced, withdrawn, or corrected, GCRI Canada shall review affected sensor records, fused records, dashboards, maps, models, Observatory records, Truth Engine records, Docket inputs, Grid inputs, Rails handoffs, and public claims.
7.3.4(h) The controlling rule shall be that sensor evidence is only as reliable as its identity, calibration, context, custody, and correction record.
7.3.5 AI-RAN and Telecommunications Signal Evidence. 7.3.5(a) AI-RAN and telecommunications signal evidence shall include evidence derived from AI-RAN systems, O-RAN systems, private wireless systems, 5G / 6G-relevant systems, non-terrestrial network interfaces, networked sensing, telecom telemetry, radio signals, environmental signals, infrastructure signals, mobility signals, edge intelligence, network observability, and mission-critical communications contexts.
7.3.5(b) AI-RAN and telecommunications signal evidence shall be recorded with signal source, network context, system boundary, owner where known, operator where applicable, provider where any, spectrum-adjacent sensitivity where material, timing, geography or safe location treatment, calibration or signal quality context, processing method, AI use where any, confidence, uncertainty, limitations, cybersecurity status, privacy status, public-safe status, and correction path.
7.3.5(c) AI-RAN and telecommunications signal evidence may support Nexus Observatory methods, remote connectivity evidence, infrastructure continuity evidence, climate and disaster evidence, public authority learning, digital twin inputs, Truth Engine methods, Docket inputs, Grid inputs, Nexus Rails handoffs, and technical baseline support.
7.3.5(d) AI-RAN and telecommunications signal evidence shall not create telecommunications regulation, spectrum authority, network control, public warning, emergency command, public authority finding, provider endorsement, service guarantee, procurement preference, certification, finance-readiness, infrastructure operation authority, or execution consequence by default.
7.3.5(e) Signal evidence shall distinguish observed signal, processed signal, inferred condition, AI-assisted interpretation, network status by another actor, provider-supplied status, public-safe summary, dashboard output, and handoff record.
7.3.5(f) Where signal evidence involves location sensitivity, critical infrastructure sensitivity, telecommunications sensitivity, public authority systems, public safety systems, emergency-adjacent communications, remote communities, or provider-supplied data, GCRI Canada shall apply heightened privacy, cybersecurity, public-safe, provider-neutrality, and public authority boundary controls.
7.3.5(g) Where signal interpretation changes because of calibration, spoof detection, sensor fusion, network configuration, provider correction, public authority concern, model update, stale data, or dispute, GCRI Canada shall correct or supersede affected records and derivative outputs.
7.3.5(h) The controlling rule shall be that telecommunications signal evidence can support observability, but it shall not make GCRI Canada the network operator, regulator, or public warning body.
7.3.6 DePIN and Distributed Infrastructure Evidence. 7.3.6(a) DePIN and distributed infrastructure evidence shall include evidence derived from distributed physical infrastructure networks, decentralized sensing systems, distributed device networks, proof infrastructure, device identity records, uptime records, location claims, hardware claims, service contribution records, proof-of-competence claims, tokenized or non-tokenized proof records, and distributed observability systems.
7.3.6(b) DePIN evidence shall be recorded with device or node identity, owner or operator where known, host where applicable, provider or protocol context where any, physical location or safe location treatment, proof method, data source, uptime method, hardware claim basis, sensor integrity, spoof risk, tamper risk, key or signing context, chain or off-chain record relationship, correction layer, confidence, uncertainty, limitations, public-safe status, and correction path.
7.3.6(c) Distributed infrastructure evidence may support observability, sensor evidence, infrastructure continuity evidence, remote connectivity evidence, climate and disaster evidence, WEFH evidence, digital twin inputs, Truth Engine comparison, Protocol Authority support, Docket inputs, Grid inputs, and public-safe reports.
7.3.6(d) DePIN or distributed infrastructure evidence shall not create official infrastructure status, public authority finding, certification, recognition, finance-readiness, provider endorsement, procurement preference, market allocation, service guarantee, public warning, emergency command, or execution consequence by default.
7.3.6(e) DePIN evidence shall distinguish device record, operator claim, protocol claim, blockchain anchor, proof receipt, observed signal, inferred condition, dashboard view, public-safe summary, and handoff record.
7.3.6(f) Where DePIN evidence includes tokens, payments, incentives, custody, financial promotion, securities, market activity, location-sensitive data, public authority data, community data, or critical infrastructure data, regulated-perimeter, privacy, cybersecurity, public-safe, and public authority boundary review shall apply.
7.3.6(g) Where device identity, location claim, uptime claim, hardware claim, sensor claim, key authority, proof method, or chain record is corrected, disputed, revoked, superseded, withdrawn, or found spoofed or tampered with, GCRI Canada shall review dependent records, dashboards, maps, evidence packs, Docket records, Grid records, Observatory records, Truth Engine records, Rails handoffs, and claims.
7.3.6(h) The controlling rule shall be that distributed infrastructure evidence is not made true by being distributed; it remains evidence only where identity, source, method, confidence, and correction are recorded.
7.3.7 Cybersecurity and Cyber-Physical Evidence. 7.3.7(a) Cybersecurity and cyber-physical evidence shall include evidence concerning cybersecurity posture, vulnerabilities, threats, incidents, logs, telemetry, intrusion indicators, vulnerability disclosures, exploit context where lawful and safe to handle, operational technology, industrial control systems, cyber-physical dependencies, supply-chain security, repository security, build pipeline security, software dependencies, key management, secrets handling, and incident learning.
7.3.7(b) Cybersecurity evidence shall be recorded with source, system or asset context, owner where known, operator where applicable, custodian, data class, sensitivity, collection method, timing, affected systems, vulnerability status, incident status, confidence, uncertainty, limitations, disclosure restrictions, public-safe status, cybersecurity controls, permitted use, prohibited use, and correction path.
7.3.7(c) Cyber-physical evidence may concern critical infrastructure, energy systems, water systems, telecommunications, ports, corridors, manufacturing, health systems, public authority systems, sensors, robotics, drones, AI systems, and other mission-critical systems where digital and physical operations interact.
7.3.7(d) Cybersecurity and cyber-physical evidence shall not create law enforcement authority, regulatory authority, public authority finding, official warning, emergency command, managed security provider status, certification, procurement approval, provider endorsement, insurance rating, finance-readiness, operational command, or execution consequence by default.
7.3.7(e) Cyber evidence shall distinguish observed log, alert, vulnerability report, threat signal, incident record, provider claim, public authority record, researcher finding, red-team result, test result, exploitability assessment, public-safe summary, and controlled disclosure.
7.3.7(f) Vulnerability and incident evidence shall be handled under controlled disclosure where public disclosure could increase harm, reveal attack paths, expose public authority systems, compromise infrastructure, harm communities, distort markets, or create public panic.
7.3.7(g) Where cyber evidence is corrected, reclassified, found false positive, found false negative, superseded, patched, mitigated, exploited, disclosed, withdrawn, or made unsafe, GCRI Canada shall review affected materials, public-safe reports, technical baselines, repositories, dashboards, Docket records, Grid records, Observatory records, Truth Engine records, Rails handoffs, and public claims.
7.3.7(h) The controlling rule shall be that cybersecurity evidence must improve safety without making GCRI Canada the regulator, responder, operator, or attacker’s guide.
7.3.8 Geospatial, Earth Observation, Satellite, Remote Sensing, and Public-Safe Mapping Evidence. 7.3.8(a) Geospatial, Earth observation, satellite, remote sensing, and public-safe mapping evidence shall include evidence derived from satellite imagery, aerial imagery, drone imagery where lawful and safe, remote sensing models, GIS layers, location intelligence, public-safe maps, environmental layers, infrastructure layers, hazard layers, community layers, and spatial analysis.
7.3.8(b) Geospatial evidence shall be recorded with source, license, collection date, processing date, spatial resolution, temporal resolution, geography, coordinate treatment where material, model or processing method, confidence, uncertainty, limitations, sensitive-location status, public-safe status, public authority status where any, community safeguard status, permitted use, prohibited use, and correction path.
7.3.8(c) Geospatial evidence may support climate evidence, disaster evidence, biodiversity evidence, WEFH evidence, infrastructure continuity evidence, remote community evidence, public authority learning, digital twin inputs, Observatory methods, Truth Engine comparison, Docket inputs, Grid inputs, and public-safe publication.
7.3.8(d) Geospatial evidence shall not create official map status, public warning, emergency command, regulatory finding, public authority decision, military or security authority, public safety directive, certification, finance-readiness, procurement approval, provider endorsement, or execution consequence by default.
7.3.8(e) Public-safe mapping shall distinguish raw imagery, processed imagery, classified imagery, inferred layer, modeled layer, scenario layer, dashboard layer, public-safe map, controlled annex map, and non-public restricted map.
7.3.8(f) Public-safe mapping controls shall protect sensitive locations, critical infrastructure, cultural sites, sacred sites, protected habitats, vulnerable communities, health-sensitive locations, public authority facilities, emergency systems, routes, ports, corridors, and protected knowledge.
7.3.8(g) Where geospatial evidence is corrected, stale, misclassified, mislocated, disputed, superseded, restricted, or withdrawn, GCRI Canada shall review affected maps, dashboards, reports, digital twins, Docket records, Grid records, Observatory records, Truth Engine records, public authority materials, community materials, and public claims.
7.3.8(h) The controlling rule shall be that maps are powerful evidence artifacts and shall be treated as public claims instruments, not neutral pictures.
7.3.9 Digital Twin, Simulation, Scenario, and Model-Based Evidence. 7.3.9(a) Digital twin, simulation, scenario, and model-based evidence shall include evidence derived from computational models, digital twins, simulations, scenario engines, stress tests, synthetic environments, system models, infrastructure models, climate models, disaster models, health-adjacent models, WEFH models, cyber models, financial-readiness input models, and other model-based methods.
7.3.9(b) Model-based evidence shall be recorded with model identity, version, purpose, owner or source, custodian, domain, data inputs, assumptions, boundary conditions, calibration status, validation status, uncertainty, sensitivity, limitations, scenario framing, output type, public-safe status, permitted use, prohibited use, and correction path.
7.3.9(c) Digital twin and scenario evidence may support public authority learning, infrastructure continuity, climate and disaster learning, WEFH learning, cybersecurity learning, Observatory methods, Truth Engine comparison, Docket inputs, Grid inputs, Nexus Risk Management, Nexus Rails, Academy materials, and public-safe reports.
7.3.9(d) Model-based evidence shall not create official prediction, public warning, emergency command, public authority decision, certification, recognition, finance-readiness, insurance rating, credit rating, public finance approval, procurement approval, provider endorsement, guarantee, operational clearance, or execution consequence by default.
7.3.9(e) Model-based records shall distinguish observed input, assumed input, inferred input, simulated condition, scenario output, forecast-like output, stress-test output, sensitivity analysis, public-safe visualization, and handoff artifact.
7.3.9(f) Model-based evidence shall avoid false precision, hidden assumptions, unstated boundaries, overgeneralization, unsupported causal claims, and presentation of scenario outputs as certain futures.
7.3.9(g) Where model assumptions, data inputs, validation status, system boundary, confidence, scenario framing, legal context, public authority context, or public-safe status changes, GCRI Canada shall review whether correction, supersession, withdrawal, downgrade, or reclassification is required.
7.3.9(h) The controlling rule shall be that models can help reason about systems, but they shall not become the systems, the decisions, or the authority.
7.3.10 Public Authority Context Evidence. 7.3.10(a) Public authority context evidence shall include evidence concerning public authority participation, public authority data, public authority records, public authority learning, public service systems, regulatory context, emergency management context, public health context, public safety context, public works context, public finance context, procurement context, and public authority-facing technology.
7.3.10(b) Public authority context evidence shall be recorded with public authority source, capacity classification, authority to contribute where applicable, official classification, permitted use, prohibited use, publication permission, reference approval, confidentiality, retention, transfer limits, correction rights, withdrawal or restriction path, non-public-authority-action status, and correction path.
7.3.10(c) Public authority context evidence may support public authority learning, public-safe reporting, Academy materials, Docket inputs, Grid inputs, Observatory methods, Truth Engine methods, technical baselines, public-good software, GRF inputs, GRA inputs, and Nexus interface instruments.
7.3.10(d) Public authority context evidence shall not create public authority delegation, endorsement, adoption, official guidance, regulatory approval, public warning, emergency command, procurement approval, funding approval, public finance approval, public-private partnership, sovereign obligation, public service decision, or public authority action by GCRI Canada.
7.3.10(e) Public authority participation shall be distinguished from public authority approval; public authority data contribution shall be distinguished from public authority adoption; public authority learning shall be distinguished from public authority delegation; and public authority-facing reports shall be distinguished from official guidance.
7.3.10(f) Names, logos, titles, quotes, photographs, agency names, jurisdiction names, attendance references, facilities, and data references shall be handled under reference controls to prevent apparent authority.
7.3.10(g) Where public authority context evidence is corrected, restricted, withdrawn, reclassified, or misdescribed, GCRI Canada shall correct affected materials, public claims, public authority references, interface records, Docket inputs, Grid inputs, Observatory records, Academy materials, and public-safe reports.
7.3.10(h) The controlling rule shall be that public authority context must be recorded precisely because public authority proximity is easily mistaken for authority.
7.3.11 Community, Indigenous, Local, Territorial, and Protected Knowledge Evidence. 7.3.11(a) Community, Indigenous, local, territorial, and protected knowledge evidence shall include knowledge, observations, records, context, lived experience, land-based knowledge, water-based knowledge, environmental knowledge, cultural knowledge, infrastructure context, service access context, hazard context, food-system knowledge, health-adjacent context, resilience knowledge, and protected knowledge contributed by or concerning communities, Indigenous or local knowledge holders, territorial actors, or rights-bearing groups.
7.3.11(b) Such evidence shall be recorded with contributor authority, community or territorial context, consent or non-consent treatment where applicable, Indigenous data sovereignty where applicable, community protocols, attribution or non-attribution, permitted use, prohibited use, publication limits, AI-use limits, data-sharing limits, public-safe mapping limits, grievance path, withdrawal or restriction path, and correction path.
7.3.11(c) Community and protected knowledge evidence may support public-safe learning, public authority learning, community safeguard methods, climate and disaster evidence, WEFH evidence, remote and northern evidence, biodiversity evidence, infrastructure evidence, Observatory methods, Docket inputs, Grid inputs, Academy materials, and public-safe reports.
7.3.11(d) Community or Indigenous participation shall not create community consent, Indigenous authorization, public authority mandate, endorsement, adoption, public warning, certification, recognition, finance-readiness, procurement approval, provider endorsement, sponsor control, or execution consequence beyond the recorded scope.
7.3.11(e) Protected knowledge shall not be treated as ordinary open data, generic evidence, sponsor-reporting content, provider-testing content, AI-training content, dashboard content, public map content, finance-facing material, or public-facing narrative merely because it is useful.
7.3.11(f) Where disclosure could expose sacred sites, cultural sites, protected habitats, vulnerable communities, sensitive locations, food sources, water sources, routes, households, knowledge holders, service gaps, or community vulnerability, GCRI Canada shall restrict, aggregate, generalize, mask, redact, delay, place in controlled annex, or refuse publication.
7.3.11(g) Where community or protected knowledge evidence is misused, over-disclosed, misattributed, incorrectly mapped, overclaimed, or used outside permitted scope, GCRI Canada shall restrict, withdraw, correct, notify affected knowledge holders where appropriate, review downstream dependencies, and preserve safeguard records.
7.3.11(h) The controlling rule shall be that community and protected knowledge evidence must preserve the people, place, sovereignty, and permissions that give it meaning.
7.3.12 Field Evidence, Operator Evidence, Host Evidence, Provider Evidence, University / Laboratory Evidence, Sponsor-Supplied Evidence, and Third-Party Evidence. 7.3.12(a) Field evidence, operator evidence, host evidence, provider evidence, university or laboratory evidence, sponsor-supplied evidence, and third-party evidence shall be treated as evidence categories requiring source-specific classification, conflict review, permission review, public-safe review, and correction path.
7.3.12(b) Field evidence shall identify field context, collector, location or safe location treatment, time period, method, conditions, safety limits, privacy limits, community safeguards, public authority context where any, and correction path.
7.3.12(c) Operator evidence shall identify owner, operator, asset, system, operational context, data source, authority to contribute, confidentiality limits, infrastructure sensitivity, operational boundary, public-safe status, non-operation status for GCRI Canada, and correction path.
7.3.12(d) Host evidence shall identify host role, facility or context, data supplied, access supplied, infrastructure supplied, staff or in-kind support where any, ownership, custody, public claims limits, confidentiality, public-safe status, and correction path.
7.3.12(e) Provider evidence shall identify provider role, technology or service, system boundary, provider claims, test conditions, data sources, conflicts, commercial sensitivity, IP restrictions, public claims limits, provider-neutrality controls, non-endorsement status, and correction path.
7.3.12(f) University and laboratory evidence shall identify research context, institutional source, method, ethics review where applicable, laboratory conditions, field conditions, data rights, IP, publication status, review status, limitations, conflicts, and correction path.
7.3.12(g) Sponsor-supplied evidence shall identify sponsor role, funding or support relationship, source, purpose, influence controls, conflicts, public claims limits, data rights, publication status, sponsor non-control status, and correction path. Sponsor-supplied evidence shall not be given higher evidentiary weight because of funding support.
7.3.12(h) Third-party evidence shall identify third-party source, authority, credibility, method where known, licensing, permission, conflicts where known, limitations, public-safe status, and whether independent corroboration is required before institutional reliance.
7.3.12(i) No field, operator, host, provider, university, laboratory, sponsor-supplied, or third-party evidence shall create certification, recognition, finance-readiness, procurement preference, provider endorsement, public authority approval, operational control, sponsor control, or execution consequence by default.
7.3.12(j) The controlling rule shall be that external evidence may enrich the Evidence Rail only when its source, incentives, limits, permissions, and correction path are visible.
7.3.13 Documentary Evidence, Policy Evidence, Legal-Context Evidence, Public Records, and Historical Evidence. 7.3.13(a) Documentary evidence, policy evidence, legal-context evidence, public records, and historical evidence shall include statutes, regulations, policies, standards, guidance, public records, reports, contracts, charters, bylaws, memoranda, interface agreements, meeting records, public authority records, historical records, institutional records, media archives where appropriately classified, and other documents relevant to GCRI Canada’s public-benefit mandate.
7.3.13(b) Documentary evidence shall be recorded with source, issuer, date, version, jurisdiction where applicable, legal or institutional status, authenticity where material, custody, access permissions, confidentiality, public-safe status, limitations, and correction or supersession path.
7.3.13(c) Policy evidence shall distinguish adopted policy, draft policy, consultation material, guidance, internal policy, external policy, public authority policy, organizational policy, and public-safe summary.
7.3.13(d) Legal-context evidence shall support understanding of legal environment, public authority context, regulatory perimeter, procurement context, public finance context, data protection context, public service context, or institutional authority, but shall not constitute legal advice by GCRI Canada unless separately authorized by competent legal professionals and records.
7.3.13(e) Public records shall be handled according to source authority, jurisdiction, publication status, official status, public-safe status, and correction or supersession path. Public records shall not create GCRI Canada authority by being referenced.
7.3.13(f) Historical evidence shall be treated as time-bound and context-bound. Historical records shall not be used as current evidence where law, technology, public authority context, institutional status, community context, or public-safe status has materially changed.
7.3.13(g) Where documentary, policy, legal-context, public record, or historical evidence is superseded, repealed, amended, corrected, withdrawn, retracted, or found unreliable, GCRI Canada shall review dependent materials and claims.
7.3.13(h) The controlling rule shall be that documents are evidence of what they validly record, not proof that the recorded state remains current or that GCRI Canada has authority over it.
7.3.14 Finance-Readiness Input Evidence Without Finance Determination. 7.3.14(a) Finance-readiness input evidence shall include technical, risk, resilience, infrastructure, climate, WEFH, cyber, AI, supply-chain, public authority learning, community safeguard, observability, benchmark, Docket, Grid, Observatory, Truth Engine, Rails, and evidence-pack materials that may support GRA or competent finance-facing actors in understanding evidence, diligence gaps, readiness gaps, or risk context.
7.3.14(b) Finance-readiness input evidence shall be recorded with source, scope, domain, intended finance-facing context, confidence, uncertainty, limitations, public-safe status, finance-safe status, permitted use, prohibited use, no-investment-advice status, no-offer status, no-solicitation status, no-underwriting status, no-rating status, no-public-finance-approval status, no-guarantee status, and correction path.
7.3.14(c) Finance-readiness input evidence may be routed to GRA, Nexus Rails, public finance readers, capital readers, insurers, lenders, underwriters, rating actors, National Companies, Project SPVs, public authorities, or other competent actors only through boundary-safe handoff records and appropriate regulated-perimeter controls.
7.3.14(d) Finance-readiness input evidence shall not create finance-readiness, investment advice, securities offering, brokerage, finder activity, placement, lending, credit approval, insurance approval, underwriting, rating, public finance approval, guarantee, capital commitment, transaction approval, fund approval, market operation, or GCRI Canada finance authority.
7.3.14(e) Finance-facing summaries shall distinguish evidence input, diligence gap, risk context, technical context, resilience context, public authority context, community safeguard context, and GRA or competent actor determination where any.
7.3.14(f) Scores, indicators, maturity-context records, resilience-context records, readiness-context records, dashboards, maps, benchmarks, or evidence packs shall not be described as ratings, guarantees, underwriting conclusions, insurance signals, credit signals, investment signals, public finance determinations, procurement rankings, or capital-readiness determinations by GCRI Canada.
7.3.14(g) Where finance-readiness input evidence is misused to imply GCRI Canada finance-readiness, investment advice, rating, underwriting, insurance approval, lending approval, public finance approval, guarantee, procurement approval, or capital commitment, GCRI Canada shall correct, restrict, withdraw, notify affected actors where appropriate, seek regulated-perimeter review where material, and preserve correction records.
7.3.14(h) The controlling rule shall be that finance-readiness input evidence may inform finance readers, but GCRI Canada shall not make finance determinations.
7.3.15 Risk, Resilience, Continuity, Infrastructure, Climate, WEFH, Health, Energy, Supply Chain, and Public Trust Evidence. 7.3.15(a) Risk, resilience, continuity, infrastructure, climate, WEFH, health, energy, supply-chain, and public trust evidence shall include evidence concerning hazard, exposure, vulnerability, resilience, adaptation, continuity, public authority learning, community resilience, infrastructure dependencies, system interdependencies, compound risk, public-safe communication, and public trust.
7.3.15(b) Risk evidence shall identify risk source, affected domain, affected actors, evidence basis, uncertainty, likelihood where applicable and appropriate, consequence where applicable and appropriate, assumptions, limitations, public-safe status, and correction path.
7.3.15(c) Resilience and continuity evidence shall identify system, function, dependency, redundancy, degraded-mode context, recovery context, community context, public authority context, infrastructure sensitivity, confidence, uncertainty, limitations, non-guarantee status, and correction path.
7.3.15(d) Climate and WEFH evidence shall identify source, geography, time period, model status where applicable, hazard context, infrastructure context, community context, public authority context, public-safe status, protected knowledge controls where applicable, confidence, uncertainty, limitations, and correction path.
7.3.15(e) Health evidence shall identify health-sensitive status, public health context where any, privacy controls, ethics review where applicable, public authority status, non-clinical-advice status, non-public-health-order status, public-safe status, and correction path.
7.3.15(f) Energy and infrastructure evidence shall identify owner where known, operator where applicable, system boundary, operational sensitivity, cyber-physical sensitivity, public authority relevance, non-operation status for GCRI Canada, non-engineering-certification status, non-warning status, public-safe status, and correction path.
7.3.15(g) Supply-chain evidence shall identify supplier or dependency context where safe and material, commercial sensitivity, strategic-material relevance, public authority relevance, export-control or sanctions sensitivity where material, competition-safety status, finance-safe status where material, and correction path.
7.3.15(h) Public trust evidence shall identify source, claim, audience, public communication context, information integrity context, reputation risk, privacy risk, source-protection status, confidence, uncertainty, limitations, public-safe status, and correction path.
7.3.15(i) These evidence categories shall not create public authority decisions, official warnings, emergency commands, insurance ratings, credit ratings, public finance approvals, regulatory determinations, procurement approvals, certifications, recognitions, provider endorsements, guarantees, or execution consequences by default.
7.3.15(j) The controlling rule shall be that risk and resilience evidence must make complex dependencies legible without becoming command, guarantee, rating, approval, or public authority action.
7.3.16 Evidence Category Register and Public-Safe Category Descriptions. 7.3.16(a) GCRI Canada shall maintain, or cause to be maintained, an Evidence Category Register identifying the evidence categories recognized by GCRI Canada, the scope of each category, applicable safeguards, permitted uses, prohibited uses, public-safe descriptions, handling controls, and correction paths.
7.3.16(b) The Evidence Category Register shall include, as applicable, scientific evidence, technical evidence, observational evidence, sensor evidence, AI-RAN and telecommunications signal evidence, DePIN and distributed infrastructure evidence, cybersecurity and cyber-physical evidence, geospatial and remote sensing evidence, digital twin and model-based evidence, public authority context evidence, community and protected knowledge evidence, field evidence, operator evidence, host evidence, provider evidence, university and laboratory evidence, sponsor-supplied evidence, third-party evidence, documentary evidence, finance-readiness input evidence, risk evidence, resilience evidence, continuity evidence, infrastructure evidence, climate evidence, WEFH evidence, health evidence, energy evidence, supply-chain evidence, and public trust evidence.
7.3.16(c) Each register entry shall identify category name, definition, included evidence types, excluded evidence types, source requirements, record requirements, classification requirements, confidence and uncertainty treatment, public-safe treatment, access controls, publication controls, interface controls, claim limits, and correction path.
7.3.16(d) Public-safe category descriptions may be published where useful to improve evidence literacy, claims discipline, public authority learning, Academy materials, public-good software documentation, public-safe reports, Docket literacy, Grid literacy, Observatory literacy, and Nexus interface clarity.
7.3.16(e) Public-safe category descriptions shall not disclose sensitive source details, protected knowledge, public authority confidential data, cyber vulnerabilities, infrastructure weaknesses, health-sensitive information, commercial secrets, export-control-sensitive information, sanctions-sensitive information, or other restricted information.
7.3.16(f) Evidence category descriptions shall not imply that a category has certification effect, recognition effect, finance-readiness effect, public authority effect, procurement effect, provider preference, sponsor control, protocol effect, or execution consequence.
7.3.16(g) Where an evidence category is corrected, split, merged, retired, superseded, restricted, or reclassified, GCRI Canada shall update the Evidence Category Register and review affected records, public materials, technical baselines, Academy materials, Docket inputs, Grid inputs, Observatory records, Truth Engine records, Rails handoffs, GRF inputs, GRA inputs, Protocol Authority support, and public claims.
7.3.16(h) The controlling rule shall be that evidence categories exist to make evidence legible and safe, not to inflate the authority of evidence by naming it.
7.4 Data-to-Evidence Rules
7.4.1 Raw Data Does Not Become Evidence Without Recorded Treatment. 7.4.1(a) Raw data shall not become evidence within GCRI Canada’s Evidence Rail merely because it has been collected, received, stored, displayed, uploaded, transmitted, measured, observed, scraped, sensed, generated, modeled, dashboarded, mapped, summarized, translated, embedded, anchored, signed, or repeated.
7.4.1(b) Raw data may be treated as evidence only where a record identifies the source, contributor, collection method, time period, geography where material and safe, data class, lawful basis or authority to contribute where applicable, custodian, custody history where material, processing history, quality treatment, classification, access controls, permitted use, prohibited use, public-safe status, and correction path.
7.4.1(c) Raw data shall remain classified as raw, preliminary, incomplete, restricted, not-for-reliance, not-for-publication, or otherwise limited until the necessary treatment records are created and reviewed.
7.4.1(d) Data treatment may include cleaning, normalization, validation, de-duplication, source comparison, quality review, permission review, privacy review, cybersecurity review, public authority review, community safeguard review, protected knowledge review, public-safe review, and correction-path creation.
7.4.1(e) No person shall convert raw data into public-facing evidence, public authority-facing evidence, finance-facing evidence, provider-facing evidence, sponsor-facing evidence, Docket input, Grid input, Observatory record, Truth Engine input, Rails handoff, Academy material, public-safe report, or technical baseline without adequate records and classification.
7.4.1(f) Raw data shall not be used to create certification, recognition, finance-readiness, procurement preference, provider endorsement, sponsor control, public authority meaning, public warning, emergency command, operational instruction, rating, guarantee, or execution consequence by implication.
7.4.1(g) Where raw data has been used beyond its recorded treatment, GCRI Canada shall review the affected use, classify the risk, correct or withdraw derivative outputs where necessary, and preserve a correction record.
7.4.1(h) The controlling rule shall be that data becomes evidence only through recorded treatment sufficient to make its source, meaning, limits, permissions, and correction path visible.
7.4.2 Telemetry Does Not Become Evidence Without Source, Context, Integrity, and Method Records. 7.4.2(a) Telemetry shall not become evidence within GCRI Canada’s Evidence Rail unless source, context, integrity, and method records establish what was measured, by what system, under what conditions, with what custody, with what limitations, and for what permitted use.
7.4.2(b) Telemetry records shall identify the telemetry source, system or device identity, owner where known, operator where applicable, host where applicable, provider where any, collection pathway, time period, update frequency, latency, system boundary, data format, transformation steps, integrity controls, access controls, and correction path.
7.4.2(c) Telemetry context shall include the operational, environmental, technical, public authority, community, infrastructure, cybersecurity, and data-governance context necessary to prevent misleading interpretation.
7.4.2(d) Telemetry integrity records shall address authenticity, completeness, continuity, clock integrity, transmission integrity, tamper risk, spoof risk, missing data, stale data, duplication, filtering, dropped records, aggregation, and processing assumptions.
7.4.2(e) Telemetry method records shall identify how telemetry is converted into evidence, including parsing, normalization, thresholding, anomaly detection, signal interpretation, model processing, AI use, human review where material, confidence treatment, uncertainty treatment, and limitation disclosure.
7.4.2(f) Telemetry shall not create public authority findings, official status, public warnings, emergency commands, infrastructure operation authority, utility direction, cybersecurity incident determination, certification, finance-readiness, procurement approval, provider endorsement, rating, guarantee, or execution consequence by default.
7.4.2(g) Where telemetry is corrected, found incomplete, spoofed, stale, misconfigured, misclassified, compromised, or improperly interpreted, GCRI Canada shall review affected records, dashboards, maps, models, Docket inputs, Grid inputs, Observatory records, Truth Engine records, Rails handoffs, public-safe reports, and public claims.
7.4.2(h) The controlling rule shall be that telemetry is not evidence until its source, context, integrity, method, and correction pathway are recorded.
7.4.3 Sensor Readings Do Not Become Evidence Without Calibration, Identity, Timing, Location, and Quality Treatment. 7.4.3(a) Sensor readings shall not become evidence within GCRI Canada’s Evidence Rail unless sensor identity, calibration, timing, location or location-treatment, data quality, custody, source context, and correction path are recorded.
7.4.3(b) Sensor identity records shall identify the sensor, device, node, owner where known, operator where applicable, host where applicable, provider where any, sensor type, firmware or configuration where material, maintenance status where available, and relationship to any Observatory node, DePIN system, AI-RAN system, dashboard, map, digital twin, or evidence pack.
7.4.3(c) Calibration records shall identify calibration status, calibration date where known, calibration method where known, reference standard where known, drift risk, measurement range, known limitations, maintenance history where available, and any reason calibration is incomplete, inferred, unavailable, or disputed.
7.4.3(d) Timing records shall identify time of measurement, time zone where material, clock source, synchronization status, latency, aggregation period, missing intervals, stale intervals, and any timing uncertainty.
7.4.3(e) Location records shall identify location, location precision, location sensitivity, safe location treatment, geospatial restrictions, public-safe mapping limits, community safeguards, infrastructure sensitivity, and any reason exact location must be restricted, generalized, masked, or withheld.
7.4.3(f) Quality treatment shall address noise, drift, missing data, outliers, tamper risk, spoof risk, environmental conditions, sensor fusion, corroboration, confidence, uncertainty, limitation, and review status.
7.4.3(g) Sensor readings shall not create official measurement status, public warning, emergency command, public authority finding, infrastructure operation authority, provider endorsement, certification, finance-readiness, procurement approval, rating, guarantee, or execution consequence by default.
7.4.3(h) Where a sensor is recalibrated, corrected, replaced, found faulty, spoofed, tampered with, mislocated, misconfigured, stale, or disputed, GCRI Canada shall review affected records, fused signals, dashboards, maps, models, Observatory records, Truth Engine records, Docket inputs, Grid inputs, Rails handoffs, and public claims.
7.4.3(i) The controlling rule shall be that a sensor reading is not evidence unless the institution can explain which sensor produced it, when, where, under what conditions, with what quality, and how it can be corrected.
7.4.4 AI-RAN Signals Do Not Become Evidence Without Signal Interpretation Methods, Confidence, Uncertainty, and Spoof / Noise Review. 7.4.4(a) AI-RAN, O-RAN, private wireless, telecommunications, non-terrestrial network, networked sensing, radio, mobility, infrastructure, environmental, and edge-intelligence signals shall not become evidence unless signal interpretation methods, source context, confidence, uncertainty, spoof review, noise review, and correction path are recorded.
7.4.4(b) AI-RAN signal records shall identify signal source, network context, system boundary, owner where known, operator where applicable, provider where any, collection method, time period, geography or safe location treatment, spectrum-adjacent sensitivity where material, device or node context where material, processing method, AI use where any, and public-safe status.
7.4.4(c) Signal interpretation methods shall identify how raw signal, processed signal, fused signal, or AI-assisted signal is converted into evidence, including method profile, assumptions, thresholds, calibration context, training or inference context where material, validation status, human review where material, and limitations.
7.4.4(d) Confidence and uncertainty treatment shall identify signal strength, completeness, quality, independence, corroboration, noise, interference, missing data, stale data, environmental effects, network configuration, model uncertainty, geospatial uncertainty, timing uncertainty, and interpretation uncertainty.
7.4.4(e) Spoof and noise review shall address intentional spoofing, unintentional noise, device misidentification, network misconfiguration, environmental interference, signal collision, model misclassification, provider data limitations, and public-safe implications.
7.4.4(f) AI-RAN and telecommunications signal evidence shall not create telecommunications regulation, spectrum authority, network control, service approval, public warning, emergency command, public authority finding, provider endorsement, procurement preference, certification, finance-readiness, infrastructure operation authority, or execution consequence by default.
7.4.4(g) Where signal interpretation changes because of spoof detection, noise review, calibration update, sensor fusion, model update, provider correction, network configuration change, stale data, public authority concern, or dispute, GCRI Canada shall correct or supersede affected records and derivative outputs.
7.4.4(h) The controlling rule shall be that AI-RAN and telecommunications signals become evidence only when interpretation, confidence, uncertainty, spoof risk, noise risk, and non-authority limits are recorded.
7.4.5 DePIN Records Do Not Become Evidence Without Device Identity, Proof Integrity, Uptime, Location, Tamper, and Method Review. 7.4.5(a) DePIN, distributed physical infrastructure, distributed device, proof infrastructure, node, uptime, location, hardware, service contribution, and proof-of-competence records shall not become evidence unless device identity, proof integrity, uptime, location or safe location treatment, tamper review, spoof review, method review, and correction path are recorded.
7.4.5(b) Device identity records shall identify device, node, hardware class, owner or operator where known, host where applicable, provider or protocol context where any, key or signing context where material, firmware or configuration where material, custody, and relationship to any sensor, Observatory node, dashboard, map, digital twin, or proof system.
7.4.5(c) Proof integrity records shall identify proof method, source record, signer or key authority, on-chain or off-chain relationship, timestamp, anchoring method where any, validation method, revocation path, supersession path, correction layer, and limitations.
7.4.5(d) Uptime records shall identify measurement method, time period, missing intervals, reporting source, independence, self-reporting risk, corroboration, confidence, uncertainty, and limitation.
7.4.5(e) Location records shall identify location claim, location proof method, location precision, safe location treatment, infrastructure sensitivity, community sensitivity, public-safe mapping limits, spoof risk, and correction path.
7.4.5(f) Tamper and spoof review shall address physical tampering, firmware tampering, key compromise, location spoofing, uptime spoofing, hardware misrepresentation, replay attack, protocol manipulation, Sybil risk where material, incentive distortion, and data-quality limitations.
7.4.5(g) DePIN records shall not create official infrastructure status, public authority finding, certification, recognition, finance-readiness, provider endorsement, procurement preference, market allocation, service guarantee, public warning, emergency command, or execution consequence by default.
7.4.5(h) Where device identity, proof integrity, uptime, location, tamper status, key authority, proof method, or chain record is corrected, disputed, revoked, superseded, withdrawn, or found spoofed or compromised, GCRI Canada shall review dependent evidence records, dashboards, maps, Docket records, Grid records, Observatory records, Truth Engine records, Rails handoffs, and public claims.
7.4.5(i) The controlling rule shall be that DePIN evidence requires proof of the proof, not merely the existence of a distributed record.
7.4.6 Cyber Logs Do Not Become Evidence Without Source, Custody, Integrity, Classification, and Cyber-Sensitive Handling. 7.4.6(a) Cyber logs, telemetry, alerts, events, vulnerability records, incident records, access records, repository logs, build pipeline logs, authentication logs, network logs, endpoint logs, operational technology logs, and cyber-physical records shall not become evidence unless source, custody, integrity, classification, cyber-sensitive handling, and correction path are recorded.
7.4.6(b) Cyber log records shall identify system source, asset context, owner where known, operator where applicable, custodian, collection method, time period, clock source, access path, retention status, completeness, chain of custody where material, and classification.
7.4.6(c) Integrity records shall address authenticity, tamper risk, missing logs, altered logs, duplicated logs, log rotation, time synchronization, retention gaps, access controls, privileged access, potential compromise, and validation where available.
7.4.6(d) Classification shall identify whether logs are public, internal, controlled, restricted, confidential, cyber-sensitive, infrastructure-sensitive, public authority-sensitive, health-sensitive, personal information, rights-bearing data, provider-sensitive, commercially sensitive, incident-sensitive, vulnerability-sensitive, or controlled-disclosure material.
7.4.6(e) Cyber-sensitive handling shall include need-to-know access, secure storage, secure transmission, controlled disclosure, redaction where needed, vulnerability disclosure controls, incident-room or evidence-room controls where applicable, public-safe publication review, and legal review where material.
7.4.6(f) Cyber logs shall not create law enforcement findings, regulatory findings, public authority findings, official warnings, emergency commands, managed security provider status, certification, insurance rating, finance-readiness, procurement approval, provider endorsement, operational command, or execution consequence by default.
7.4.6(g) Where cyber logs are corrected, found incomplete, found false positive, found false negative, compromised, superseded, reclassified, disclosed, withdrawn, or made unsafe, GCRI Canada shall review affected incident records, vulnerability records, public-safe reports, technical baselines, repositories, dashboards, Docket records, Grid records, Observatory records, Truth Engine records, Rails handoffs, and public claims.
7.4.6(h) The controlling rule shall be that cyber logs become evidence only when they are handled as sensitive technical records, not as raw material for unsafe conclusions.
7.4.7 Digital Twin Outputs Do Not Become Evidence Without Assumption, Input, Calibration, Validation, and Limitation Records. 7.4.7(a) Digital twin outputs, simulation outputs, scenario outputs, model-based outputs, stress-test outputs, system-model outputs, infrastructure-model outputs, climate-model outputs, disaster-model outputs, cyber-model outputs, WEFH-model outputs, and related computational outputs shall not become evidence unless assumption, input, calibration, validation, limitation, uncertainty, and correction records are created.
7.4.7(b) Assumption records shall identify model assumptions, boundary conditions, scenario framing, excluded conditions, simplifications, dependencies, parameter choices, causal assumptions where any, time horizon, geography, and affected system boundaries.
7.4.7(c) Input records shall identify data sources, input quality, input date, input geography, input restrictions, public authority data status where any, community or protected knowledge status where any, provider-supplied data where any, sponsor-supported data where any, and transformation history.
7.4.7(d) Calibration records shall identify calibration method, calibration data, calibration date, calibration limits, domain relevance, model fit, drift risk, and uncertainty.
7.4.7(e) Validation records shall identify validation method, reviewer, test data where any, benchmark or comparison where any, validation conditions, validation scope, failures, limitations, and whether validation is internal, external, preliminary, controlled, public-safe, or unsuitable for public claim.
7.4.7(f) Limitation records shall identify what the model does not show, what conditions are outside scope, where false precision risk exists, where causal claims are unsupported, where generalization is unsafe, where public-safe publication is restricted, and what downstream uses require further review.
7.4.7(g) Digital twin and model outputs shall not create official forecasts, public warnings, emergency commands, public authority decisions, certification, recognition, finance-readiness, insurance ratings, credit ratings, public finance approvals, procurement approvals, provider endorsements, guarantees, operational clearances, or execution consequences by default.
7.4.7(h) Where assumptions, inputs, calibration, validation, model version, legal context, public authority context, public-safe status, or evidence basis changes, GCRI Canada shall review whether correction, supersession, withdrawal, downgrade, or reclassification is required.
7.4.7(i) The controlling rule shall be that model outputs become evidence only when the institution can see the assumptions beneath them and the limits around them.
7.4.8 AI Outputs Do Not Become Evidence Without Model, Prompt / Input, Retrieval, Data Source, Human Review Where Material, and Inference Records. 7.4.8(a) AI outputs, machine learning outputs, foundation model outputs, generative AI outputs, agentic AI outputs, automated classifications, AI-assisted summaries, AI-assisted evidence routing, AI-assisted source comparison, AI-assisted translation, AI-assisted public-safe review, and AI-assisted dashboard outputs shall not become evidence unless model, prompt or input, retrieval, data source, human review where material, inference, limitation, and correction records are created.
7.4.8(b) Model records shall identify model or tool used where material, version where known, provider where any, deployment context, purpose, known limitations, data-processing terms where material, retention status where material, training or fine-tuning restrictions, embedding restrictions, retrieval restrictions, and public-safe constraints.
7.4.8(c) Prompt or input records shall identify input category, source materials, instructions, constraints, user role where material, system context where material, sensitive data status, public authority data status, community or protected knowledge status, and permitted use.
7.4.8(d) Retrieval and data source records shall identify source repositories, documents, datasets, APIs, knowledge bases, embeddings, search results, citations where applicable, data freshness, source restrictions, and any known gaps or conflicts.
7.4.8(e) Human review shall be required where AI outputs are material to public-facing claims, public authority-facing materials, finance-facing materials, procurement-facing materials, provider-facing materials, sponsor-facing materials, community-facing materials, rights-bearing data, protected knowledge, cybersecurity, health-sensitive contexts, infrastructure-sensitive contexts, or other high-risk use.
7.4.8(f) Inference records shall identify AI output category, confidence where available and meaningful, uncertainty, hallucination risk, bias or drift risk, source-grounding status, human review status, limitations, public-safe status, permitted use, prohibited use, and correction path.
7.4.8(g) AI outputs shall not create authority, certification, recognition, finance-readiness, investment advice, insurance approval, underwriting conclusion, rating, public authority decision, public warning, emergency command, procurement approval, provider endorsement, professional advice, legal advice, clinical advice, operational command, or execution consequence by default.
7.4.8(h) Where AI outputs are found hallucinated, biased, unsafe, stale, unsupported, source-misaligned, privacy-invasive, overbroad, misleading, or misused, GCRI Canada shall correct, withdraw, re-run with safeguards where appropriate, conduct human review, review downstream dependencies, and preserve correction records.
7.4.8(i) The controlling rule shall be that AI output is not evidence merely because it is fluent, confident, fast, or useful; it becomes evidence only through source-grounded, reviewed, bounded, and correctable records.
7.4.9 Public Authority Inputs Do Not Become Public Authority Decisions by Evidence Treatment. 7.4.9(a) Public authority inputs shall not become public authority decisions, official guidance, regulatory findings, public warnings, emergency commands, procurement approvals, funding approvals, public finance approvals, public service decisions, eligibility determinations, entitlement determinations, public health orders, public safety directives, public works directions, or sovereign obligations merely because GCRI Canada receives, structures, reviews, publishes, routes, dashboards, maps, summarizes, or treats such inputs as evidence.
7.4.9(b) Public authority input records shall identify public authority source, capacity classification, authority to contribute where applicable, official classification, permitted use, prohibited use, publication permission, reference approval, confidentiality, retention, transfer limits, correction rights, withdrawal or restriction path, and non-public-authority-action status.
7.4.9(c) Public authority attendance, data contribution, quote, funding, venue, logo, title, photograph, review, participation, or receipt of GCRI Canada materials shall not create endorsement, adoption, official guidance, public authority delegation, public-private partnership, procurement approval, public finance approval, public warning, emergency command, or public authority decision by default.
7.4.9(d) Public authority inputs may support public authority learning, technical literacy, evidence literacy, public-safe reports, controlled annexes, Academy materials, Docket inputs, Grid inputs, Observatory methods, Truth Engine methods, Rails handoffs, GRF inputs, GRA inputs, or Protocol Authority support only within recorded boundary language.
7.4.9(e) Public authority-facing outputs shall include non-delegation, non-endorsement, non-official-guidance, non-public-authority-decision, non-warning, non-command, non-regulatory, non-procurement, non-public-finance-approval, non-public-service-operation, and correction language where material.
7.4.9(f) Where public authority inputs are misdescribed as decisions, approvals, endorsements, official guidance, warnings, commands, funding approvals, procurement approvals, public finance approvals, or public service actions, GCRI Canada shall correct, restrict, withdraw, notify the relevant public authority where appropriate, seek public authority boundary review where material, and preserve correction records.
7.4.9(g) Where public authority capacity is ambiguous, GCRI Canada shall treat the input under the most restrictive non-authority interpretation until the competent public authority clarifies the record.
7.4.9(h) The controlling rule shall be that public authority evidence can inform GCRI Canada’s records, but only public authorities acting through their own lawful processes make public authority decisions.
7.4.10 Community and Indigenous Inputs Require Context, Authority, Consent / Non-Consent Where Applicable, Protected Knowledge Treatment, and Public-Safe Handling. 7.4.10(a) Community, Indigenous, local, territorial, cultural, environmental, land-based, water-based, food-system, health-adjacent, resilience, service-access, infrastructure, and protected knowledge inputs shall not become evidence unless context, contributor authority, consent or non-consent treatment where applicable, protected knowledge treatment, public-safe handling, and correction path are recorded.
7.4.10(b) Community and Indigenous input records shall identify contributor authority, community or territorial context, knowledge type, consent or non-consent treatment where applicable, Indigenous data sovereignty where applicable, community protocols, cultural protocols, attribution or non-attribution, permitted use, prohibited use, AI-use limits, data-sharing limits, publication limits, public-safe mapping limits, grievance path, withdrawal or restriction path, and correction path.
7.4.10(c) Community and Indigenous inputs shall not be treated as ordinary open data, generic evidence, sponsor-reporting content, provider-testing content, AI-training content, dashboard content, public map content, finance-facing material, media narrative, or public-facing story merely because they are useful or compelling.
7.4.10(d) Community participation, knowledge contribution, attendance, hosting, public statement, photograph, quotation, data contribution, or review shall not create community consent, Indigenous authorization, public authority mandate, endorsement, adoption, certification, recognition, finance-readiness, procurement approval, provider endorsement, sponsor control, public warning, emergency command, or execution consequence beyond the recorded scope.
7.4.10(e) Public-safe handling shall protect sacred sites, cultural sites, protected habitats, vulnerable communities, sensitive locations, food sources, water sources, routes, households, knowledge holders, community vulnerabilities, small-population identifiability, public authority relationships, and protected knowledge from unsafe disclosure.
7.4.10(f) Where disclosure, mapping, dashboarding, modeling, AI use, translation, summarization, handoff, or publication may create community harm, GCRI Canada shall restrict, aggregate, generalize, mask, redact, delay, localize, compute-to-data, place in controlled annex, refuse publication, or withdraw.
7.4.10(g) Where community or Indigenous inputs are misused, over-disclosed, misattributed, incorrectly mapped, overclaimed, used outside permitted scope, or rendered unsafe, GCRI Canada shall restrict, withdraw, correct, notify affected knowledge holders where appropriate and lawful, review downstream dependencies, and preserve safeguard records.
7.4.10(h) The controlling rule shall be that community and Indigenous inputs become evidence only when the authority, context, permissions, protections, and public-safe limits that give them meaning remain attached.
7.4.11 Provider or Sponsor Inputs Require Conflict Review, Independence Treatment, and Claims Controls. 7.4.11(a) Provider inputs, vendor inputs, supplier inputs, sponsor inputs, donor inputs, funder inputs, host-supported inputs, capital-reader inputs, and other interested-party inputs shall not become institutional evidence unless conflict review, independence treatment, source treatment, claims controls, and correction path are recorded.
7.4.11(b) Provider input records shall identify provider role, technology or service, system boundary, input type, evidence supplied, test conditions where any, data supplied, IP restrictions, commercial sensitivity, conflicts, public claims limits, provider-neutrality status, non-endorsement status, non-procurement status, and correction path.
7.4.11(c) Sponsor input records shall identify sponsor role, funding or support relationship, input type, purpose, influence controls, conflicts, data rights, publication restrictions, public claims limits, sponsor non-control status, and correction path.
7.4.11(d) Conflict review shall address financial interest, commercial interest, procurement interest, market-position interest, reputational interest, policy interest, public authority interest, sponsor interest, provider interest, host interest, and any relationship that may affect evidentiary weight, public claims, or public trust.
7.4.11(e) Independence treatment shall distinguish independent evidence, internally generated evidence, provider-supplied evidence, sponsor-supplied evidence, host-supplied evidence, public authority-supplied evidence, university-supplied evidence, community-supplied evidence, and third-party evidence.
7.4.11(f) Provider or sponsor inputs shall not be given higher evidentiary weight because of funding, sponsorship, participation, access, market prominence, institutional relationship, public authority proximity, media attention, or Nexus proximity.
7.4.11(g) Claims controls shall prohibit use of provider or sponsor inputs to imply endorsement, preferred provider status, procurement preference, certification, recognition, finance-readiness, public authority approval, market superiority, sponsor control, or execution consequence by default.
7.4.11(h) Where provider or sponsor inputs are corrected, disputed, found misleading, found incomplete, found conflict-affected, withdrawn, restricted, or overclaimed, GCRI Canada shall review dependent materials, public claims, Docket inputs, Grid inputs, Observatory records, Truth Engine records, Rails handoffs, public authority materials, finance-facing materials, procurement-facing materials, and provider or sponsor references.
7.4.11(i) The controlling rule shall be that interested-party inputs may be useful evidence only when interests, limits, independence, and claim boundaries are visible.
7.4.12 Data-to-Evidence Conversion Requires Method, Record, Classification, Custody, Review, Limitation, and Correction Path. 7.4.12(a) Data-to-evidence conversion shall require method, record, classification, custody, review, limitation, and correction path. No data, telemetry, sensor reading, AI-RAN signal, DePIN record, cyber log, digital twin output, AI output, public authority input, community input, provider input, sponsor input, dashboard view, map layer, blockchain anchor, proof receipt, document, statement, or public communication shall become institutional evidence without such conversion discipline.
7.4.12(b) Method shall identify how the input is treated, processed, interpreted, validated, compared, classified, reviewed, summarized, translated, modeled, dashboarded, mapped, or converted into an evidence record.
7.4.12(c) Record shall identify source, contributor, time, domain, version, evidence class, data class, authority to contribute where applicable, permitted use, prohibited use, public-safe status, and correction path.
7.4.12(d) Classification shall identify sensitivity, public-safe status, public authority status, privacy status, cybersecurity status, sovereign data status, community or protected knowledge status, finance-facing status, procurement-facing status, provider-facing status, sponsor-facing status, and any restricted handling requirements.
7.4.12(e) Custody shall identify who holds, controls, processes, stores, edits, transmits, publishes, routes, archives, or may correct the input or evidence record, including owner, custodian, steward, processor, repository, data room, controlled room, API, dashboard, or interface where material.
7.4.12(f) Review shall address source integrity, method integrity, data rights, privacy, cybersecurity, sovereignty, public authority limits, community safeguards, protected knowledge, conflicts, public-safe publication, finance boundaries, procurement boundaries, provider neutrality, sponsor non-control, and correction needs.
7.4.12(g) Limitation shall identify what the evidence does not show, what uses are prohibited, what confidence or uncertainty applies, what authority is not created, what public claims are not permitted, what downstream uses require further review, and what conditions may trigger correction or withdrawal.
7.4.12(h) Correction path shall identify how the evidence may be corrected, superseded, withdrawn, retracted, downgraded, restricted, reclassified, reinstated, archived, and reviewed for downstream dependencies.
7.4.12(i) Data-to-evidence conversion shall not create certification, recognition, finance-readiness, public authority effect, public warning, emergency command, procurement approval, provider endorsement, sponsor control, rating, guarantee, operational authority, protocol effect, legal status, or execution consequence by default.
7.4.12(j) The controlling rule shall be that evidence is not a data type; it is a disciplined institutional state created by method, record, classification, custody, review, limitation, and correctionability.
7.5 Source Lineage, Provenance, and Custody
7.5.1 Source Lineage as Mandatory Evidence Attribute. 7.5.1(a) Source lineage shall be a mandatory attribute of evidence within the Evidence Rail. No evidence shall be treated as institutionally valid, decision-grade, public-safe, routeable, publishable, handoff-ready, Docket-ready, Grid-ready, Observatory-ready, Truth Engine-ready, Rails-ready, GRF-input-ready, GRA-input-ready, Protocol Authority-supporting, or externally claimable unless its source lineage is recorded to a degree proportionate to its intended use and risk.
7.5.1(b) Source lineage shall identify the origin of the evidence, the person, institution, system, device, model, repository, public authority, provider, sponsor, host, community, knowledge holder, dataset, sensor, software asset, dashboard, map, API, document, or record from which the evidence arose, and the pathway by which the evidence entered GCRI Canada’s custody, awareness, repository, workflow, or interface.
7.5.1(c) Source lineage shall preserve the distinction between original source, intermediate source, transmitting source, summarizing source, translating source, transforming source, AI-assisted source, model-generated source, dashboard-displayed source, map-displayed source, public authority source, provider source, sponsor source, community source, and publication source.
7.5.1(d) Source lineage shall be recorded for raw data, processed data, telemetry, sensor readings, AI-RAN signals, DePIN records, cyber logs, model outputs, AI outputs, geospatial layers, digital twin outputs, public authority inputs, community inputs, provider inputs, sponsor inputs, documentary materials, public records, field observations, benchmark results, pilot records, demonstration records, validation sprint records, and derived evidence.
7.5.1(e) Evidence lacking adequate source lineage shall be classified as incomplete, preliminary, restricted, not-for-reliance, not-for-publication, source-unverified, or otherwise limited until source lineage is established or the evidence is withdrawn, rejected, quarantined, or used only for an appropriately bounded purpose.
7.5.1(f) Source lineage shall not by itself create reliability, authority, legality, public-safe status, finance-readiness, certification, recognition, procurement preference, public authority meaning, provider endorsement, sponsor control, or execution consequence. Source lineage establishes traceability; the evidentiary weight of the source shall require further review.
7.5.1(g) Where source lineage is missing, broken, disputed, altered, unverifiable, inconsistent, incomplete, stale, circular, or dependent on unsupported assertions, GCRI Canada shall record the deficiency, limit use, prohibit inflated claims, and determine whether correction, restriction, further review, or withdrawal is required.
7.5.1(h) The controlling rule shall be that evidence without source lineage cannot carry institutional trust beyond the narrowest safe use.
7.5.2 Provenance as Evidence Integrity Requirement. 7.5.2(a) Provenance shall be an evidence integrity requirement for all material evidence within the Evidence Rail. Provenance shall identify how evidence came into being, how it was collected, generated, observed, measured, modeled, processed, transformed, reviewed, transferred, stored, displayed, summarized, published, corrected, or superseded.
7.5.2(b) Provenance records shall identify original creation context, collection method, generation method, instrument or system used, human contributor where known, machine contributor where any, model or software used where material, processing steps, transformations, translations, aggregations, redactions, de-identification, masking, public-safe edits, and derivative outputs.
7.5.2(c) Provenance shall apply to documents, datasets, code, public-good software, technical baselines, APIs, schemas, dashboards, maps, digital twins, AI outputs, model outputs, sensor records, blockchain anchors, proof receipts, DePIN records, cyber logs, geospatial layers, public authority materials, community materials, provider materials, sponsor materials, and public-safe publications.
7.5.2(d) Provenance shall distinguish original evidence from copies, excerpts, summaries, translations, screenshots, visualizations, dashboard views, model outputs, AI-assisted summaries, public-safe summaries, controlled annexes, media descriptions, and public-facing claims.
7.5.2(e) Provenance records shall identify whether the evidence has been altered, cleaned, normalized, enriched, fused, compressed, anonymized, redacted, filtered, mapped, embedded, vectorized, modeled, reclassified, corrected, superseded, withdrawn, retracted, or archived.
7.5.2(f) Cryptographic signatures, hashes, blockchain anchors, timestamps, content credentials, watermarking, provenance metadata, or proof receipts may support provenance but shall not be conclusive by default. Technical proof of existence, custody, or integrity shall not by itself prove accuracy, authority, consent, public-safe status, legality, or institutional validity.
7.5.2(g) Where provenance is unclear, inconsistent, machine-obscured, tool-dependent, provider-controlled, sponsor-framed, public authority-sensitive, community-sensitive, or legally restricted, GCRI Canada shall classify the evidence accordingly and restrict claims until adequate review is completed.
7.5.2(h) The controlling rule shall be that provenance protects evidence from becoming detached from the process that made it meaningful.
7.5.3 Custody as Chain-of-Handling Requirement. 7.5.3(a) Custody shall be a chain-of-handling requirement for evidence within the Evidence Rail. Custody records shall identify who or what held, controlled, accessed, processed, transferred, altered, reviewed, approved, published, routed, corrected, restricted, withdrawn, retracted, or archived the evidence.
7.5.3(b) Custody may include physical custody, digital custody, repository custody, data-room custody, clean-room custody, controlled-room custody, API custody, dashboard custody, model-environment custody, compute-environment custody, public authority custody, provider custody, sponsor custody, host custody, community custody, and archival custody.
7.5.3(c) Custody records shall identify owner where known, custodian, steward, processor, maintainer, reviewer, approver where applicable, access roles, transfer path, storage location, system of record, repository, version, logs, access controls, security controls, retention status, and correction path.
7.5.3(d) Custody shall be especially strict for evidence involving rights-bearing data, public authority data, health-sensitive data, infrastructure-sensitive data, cyber-sensitive data, sovereign data, Indigenous or local knowledge, protected knowledge, commercial secrets, export-control-sensitive information, sanctions-sensitive information, controlled technology, finance-facing materials, procurement-facing materials, and public-safe restricted materials.
7.5.3(e) A break, gap, uncertainty, unexplained alteration, unauthorized access, uncontrolled copy, unsanctioned export, unlogged transfer, lost record, or disputed custody event shall be recorded and shall trigger review proportionate to the sensitivity and intended use of the evidence.
7.5.3(f) Custody shall not create authority by possession. GCRI Canada’s custody of evidence shall not convert the evidence into GCRI Canada ownership, public authority status, finance-readiness, certification, recognition, procurement approval, provider endorsement, sponsor approval, operational command, public warning, or execution consequence.
7.5.3(g) Where custody risk affects evidence reliability, confidentiality, public-safe status, privacy, cybersecurity, source protection, or legal compliance, GCRI Canada shall restrict use, correct records, notify affected actors where appropriate, seek review where material, and preserve incident and correction records.
7.5.3(h) The controlling rule shall be that evidence must have a traceable chain of handling before it can have a trustworthy institutional life.
7.5.4 Timestamp, Version, Location, System, Contributor, Authority, and Collection Context. 7.5.4(a) Evidence records shall include timestamp, version, location or safe location treatment, system identity, contributor identity or contributor class where appropriate, source authority, and collection context sufficient to support interpretation, review, public-safe handling, and correction.
7.5.4(b) Timestamp records shall identify creation time, collection time, observation time, receipt time, processing time, publication time, review time, correction time, supersession time, withdrawal time, retraction time, or archival time where material, including time zone, clock source, synchronization status, latency, aggregation period, and uncertainty where relevant.
7.5.4(c) Version records shall identify dataset version, model version, software version, API version, schema version, technical baseline version, dashboard version, map version, report version, benchmark version, pilot record version, demonstration record version, public-safe summary version, and correction version where material.
7.5.4(d) Location records shall identify physical location, jurisdiction, geography, facility, node, sensor location, public authority jurisdiction, community context, safe location treatment, location precision, location uncertainty, public-safe mapping restrictions, infrastructure sensitivity, community sensitivity, and protected knowledge limits where material.
7.5.4(e) System records shall identify the system, device, software, model, API, repository, dashboard, map, data room, compute environment, sensor, AI-RAN system, DePIN node, Observatory node, public authority system, provider system, host system, or source system from which evidence arose or through which it was processed.
7.5.4(f) Contributor records shall identify contributor, contributor role, contributor authority, contributor affiliation where material, conflicts where known, capacity classification where a public authority is involved, provider status where a provider is involved, sponsor status where a sponsor is involved, and community or knowledge-holder authority where protected knowledge is involved.
7.5.4(g) Collection context shall identify purpose, method, conditions, audience where material, constraints, permissions, legal basis where applicable, public authority context, community context, operational context, finance-facing context, public-safe context, and known limitations.
7.5.4(h) The controlling rule shall be that evidence must be located in time, version, place, system, contributor, authority, and context before it can be responsibly interpreted.
7.5.5 Permissions, Licenses, Consent / Non-Consent Where Applicable, Public Authority Authority, Data Sharing Terms, and Use Limits. 7.5.5(a) Evidence shall include permission, license, consent or non-consent where applicable, public authority authority, data-sharing terms, confidentiality terms, IP terms, AI-use terms, publication terms, transfer terms, retention terms, and use limits sufficient to determine whether GCRI Canada may receive, process, retain, analyze, model, dashboard, map, publish, route, share, train on, embed, archive, or correct the evidence.
7.5.5(b) Permission records shall identify the source of permission, scope, permitted uses, prohibited uses, duration, revocation or withdrawal rights, publication rights, sublicensing rights where any, onward-sharing restrictions, public-safe restrictions, and correction obligations.
7.5.5(c) License records shall identify copyright, database rights, software license, open-source license, data license, public authority license, provider license, university or laboratory license, community license or protocol where applicable, attribution obligations, non-commercial limitations, share-alike obligations, field-of-use restrictions, confidentiality limits, and termination conditions.
7.5.5(d) Consent and non-consent treatment shall be recorded where evidence involves personal information, rights-bearing data, health-sensitive data, community knowledge, Indigenous or local knowledge, protected knowledge, identifiable persons, vulnerable persons, public-facing participation, or other contexts requiring consent, non-consent recognition, refusal, withdrawal, or restriction.
7.5.5(e) Public authority authority records shall identify whether the public authority has authority to contribute, share, authorize, publish, rely upon, or restrict the evidence, and shall preserve capacity classification, official classification, permitted use, prohibited use, publication permission, reference approval, confidentiality, retention, transfer limits, correction rights, and withdrawal or restriction path.
7.5.5(f) Data-sharing terms shall identify data controller or equivalent actor where applicable, processor or equivalent actor where applicable, lawful basis where applicable, hosting location, security controls, cross-border transfer, data residency, compute-to-data requirements, AI-use limits, access limits, retention, deletion, breach handling, and correction obligations.
7.5.5(g) Use limits shall travel with evidence through derivative records, summaries, dashboards, maps, reports, Academy materials, Docket records, Grid records, Observatory records, Truth Engine records, Rails handoffs, GRF inputs, GRA inputs, Protocol Authority support, public authority handoffs, provider-facing materials, sponsor-facing materials, and archives.
7.5.5(h) The controlling rule shall be that evidence cannot be more usable than its permissions, licenses, authority, consent status, data-sharing terms, and use limits allow.
7.5.6 Source Reliability, Source Bias, Source Independence, Source Conflict, and Source Completeness. 7.5.6(a) GCRI Canada shall assess source reliability, source bias, source independence, source conflict, and source completeness for material evidence, proportionate to intended use, sensitivity, public-facing status, public authority relevance, finance-facing relevance, procurement-facing relevance, provider relevance, sponsor relevance, community relevance, and public-safe risk.
7.5.6(b) Source reliability review shall consider source identity, expertise, institutional role, collection method, track record, documentation quality, reproducibility, corroboration, timeliness, completeness, error history, correction history, and whether the source is primary, secondary, derived, inferred, or promotional.
7.5.6(c) Source bias review shall consider incentives, perspective, selection effects, measurement bias, reporting bias, survivorship bias, publication bias, political bias, commercial bias, sponsor bias, provider bias, public authority bias, media framing, model bias, data bias, community under-representation, and historical under-service.
7.5.6(d) Source independence review shall distinguish independent sources, mutually dependent sources, circular sources, common-origin sources, shared-data sources, provider-supplied sources, sponsor-supplied sources, public authority sources, community sources, media sources, and machine-generated sources.
7.5.6(e) Source conflict review shall identify financial interests, commercial interests, procurement interests, market-position interests, reputation interests, funding relationships, sponsor relationships, provider relationships, host relationships, public authority relationships, political or policy interests, academic conflicts, and personal conflicts where material.
7.5.6(f) Source completeness review shall identify missing data, missing populations, missing geographies, missing time periods, missing failure cases, missing negative results, missing uncertainty, missing affected communities, missing public authority context, missing provider context, missing sponsor context, missing limitations, and missing correction path.
7.5.6(g) Source reliability shall not be presumed from prestige, public authority status, market prominence, sponsor support, provider scale, media coverage, technical sophistication, blockchain anchoring, AI fluency, dashboard polish, or Nexus proximity.
7.5.6(h) Where source reliability is uncertain, bias is material, independence is weak, conflicts exist, or completeness is limited, GCRI Canada shall classify the evidence accordingly, limit claims, require corroboration where appropriate, and preserve correction triggers.
7.5.6(i) The controlling rule shall be that source quality is not a binary attribute; it is a recorded judgment that must travel with the evidence.
7.5.7 Provider-Supplied and Sponsor-Supplied Source Treatment. 7.5.7(a) Provider-supplied and sponsor-supplied sources shall be treated as interested-party sources requiring source-lineage review, conflict review, independence treatment, evidentiary-weight treatment, claims controls, provider-neutrality controls, sponsor non-control controls, competition safety review where material, and correction path.
7.5.7(b) Provider-supplied source records shall identify the provider, technology, service, product, system boundary, data supplied, method supplied, test conditions where any, configuration, version, commercial sensitivity, IP restrictions, conflict interests, claims requested or made, and public claims limits.
7.5.7(c) Sponsor-supplied source records shall identify the sponsor, donor, funder, or supporter; funding or support relationship; purpose of source provision; influence controls; restrictions; conflicts; data rights; public claims limits; and sponsor non-control status.
7.5.7(d) Provider-supplied or sponsor-supplied evidence shall not be excluded solely because it is interested-party evidence, but it shall not be treated as independent evidence unless independence is separately recorded.
7.5.7(e) Provider or sponsor access to the Evidence Rail shall not permit alteration, suppression, selective disclosure, classification pressure, public-safe manipulation, benchmark manipulation, dashboard framing, report framing, or public claims beyond the record.
7.5.7(f) Provider-supplied benchmark data, performance data, technical documentation, demonstrations, pilots, validation sprint materials, sensor records, AI outputs, cybersecurity records, DePIN records, or software records shall be reviewed for test conditions, reproducibility where appropriate, conflicts, missing limitations, public-safe status, and prohibited claims.
7.5.7(g) Sponsor-supplied narratives, impact claims, public benefit claims, finance-facing claims, public authority claims, community claims, or public trust claims shall be treated as claims requiring verification, not evidence of the truth asserted by default.
7.5.7(h) No provider-supplied or sponsor-supplied source shall create provider endorsement, sponsor control, procurement preference, certification, recognition, finance-readiness, public authority approval, market allocation, rating, guarantee, public-safe status, or execution consequence by default.
7.5.7(i) Where provider-supplied or sponsor-supplied source material is corrected, disputed, withdrawn, found misleading, found incomplete, conflict-affected, unsafe, or overclaimed, GCRI Canada shall review dependent materials, public claims, handoffs, dashboards, Docket inputs, Grid inputs, Observatory records, Truth Engine records, Rails records, and interface references.
7.5.7(j) The controlling rule shall be that provider and sponsor sources may inform evidence, but they shall never own the meaning of evidence.
7.5.8 Public Authority Source Treatment. 7.5.8(a) Public authority sources shall be treated with source-lineage, capacity, authority, confidentiality, public-safe, and non-delegation discipline. Public authority source status may increase relevance, but it shall not create GCRI Canada public authority, public authority approval, official guidance, or public authority decision by default.
7.5.8(b) Public authority source records shall identify the public authority, office, agency, department, Crown entity, municipality, Indigenous government or authority, provincial or territorial authority, federal body, regulator, emergency management actor, public health actor, public safety actor, public works actor, public finance actor, procurement actor, or other public actor supplying or referenced by the source.
7.5.8(c) Capacity classification shall identify whether the public authority source is acting as public record issuer, data contributor, observer, learner, technical participant, reviewer, host, funder, regulator attending in non-regulatory capacity, emergency management learner, public finance reader, procurement observer, or official decision-maker acting only through a separate public authority process.
7.5.8(d) Public authority authority records shall identify authority to contribute, authority to share, authority to authorize publication, official classification, legal restrictions, confidentiality, permitted use, prohibited use, retention, transfer limits, reference approval, correction rights, and withdrawal or restriction path.
7.5.8(e) Public authority sources shall be distinguished from public authority decisions. A public authority source may provide context, data, record, view, or learning input without issuing endorsement, adoption, official guidance, regulatory approval, public warning, emergency command, procurement approval, funding approval, public finance approval, or sovereign obligation.
7.5.8(f) Public authority names, logos, titles, quotes, photographs, agency names, jurisdiction names, attendance, facilities, data references, and funding references shall be handled under reference controls to prevent apparent authority.
7.5.8(g) Where a public authority source is corrected, withdrawn, restricted, reclassified, superseded, misattributed, or overclaimed, GCRI Canada shall correct affected materials, public claims, public authority references, interface records, Docket inputs, Grid inputs, Observatory records, Truth Engine records, Academy materials, and public-safe reports.
7.5.8(h) The controlling rule shall be that public authority sources are important evidence sources, but only public authorities acting through lawful processes create public authority acts.
7.5.9 Community and Protected Knowledge Source Treatment. 7.5.9(a) Community, Indigenous, local, territorial, and protected knowledge sources shall be treated with context, contributor authority, consent or non-consent treatment where applicable, Indigenous data sovereignty where applicable, community protocol, protected knowledge, public-safe, grievance, withdrawal, and correction discipline.
7.5.9(b) Community and protected knowledge source records shall identify contributor authority, community or territorial context, knowledge type, collection context, consent or non-consent treatment where applicable, Indigenous data sovereignty where applicable, community protocols, cultural protocols, attribution or non-attribution, permitted use, prohibited use, publication limits, AI-use limits, data-sharing limits, public-safe mapping limits, grievance path, withdrawal or restriction path, and correction path.
7.5.9(c) Community or protected knowledge shall not be treated as ordinary open data, generic public evidence, sponsor-reporting content, provider-testing content, AI-training content, dashboard content, map content, finance-facing material, media narrative, or public-facing story merely because it is useful, vivid, or persuasive.
7.5.9(d) Source treatment shall protect sacred sites, cultural sites, protected habitats, vulnerable communities, sensitive locations, food sources, water sources, routes, households, knowledge holders, small-population identifiability, public authority relationships, community vulnerability, and protected knowledge from unsafe disclosure or extraction.
7.5.9(e) Community participation, Indigenous participation, local knowledge contribution, attendance, hosting, quotation, public statement, photograph, data contribution, or review shall not create community consent, Indigenous authorization, endorsement, adoption, public authority mandate, certification, recognition, finance-readiness, procurement approval, provider endorsement, sponsor control, public warning, emergency command, or execution consequence beyond the recorded scope.
7.5.9(f) Where source restrictions require aggregation, generalization, masking, redaction, delayed release, controlled annex treatment, localization, compute-to-data, or non-public handling, those restrictions shall travel with all derivative evidence and outputs.
7.5.9(g) Where community or protected knowledge source material is misused, over-disclosed, misattributed, incorrectly mapped, overclaimed, used outside permitted scope, or rendered unsafe, GCRI Canada shall restrict, withdraw, correct, notify affected knowledge holders where appropriate and lawful, review downstream dependencies, and preserve safeguard records.
7.5.9(h) The controlling rule shall be that community and protected knowledge sources remain connected to the people, places, permissions, and protections that give them meaning.
7.5.10 Machine-Generated Source Treatment. 7.5.10(a) Machine-generated sources shall include AI outputs, machine learning outputs, foundation model outputs, agentic AI outputs, automated classifications, sensor outputs, telemetry streams, AI-RAN signals, DePIN records, blockchain records, proof receipts, smart contract events, digital twin outputs, simulation outputs, dashboard outputs, API outputs, system logs, cyber logs, and other records generated in whole or in part by technical systems.
7.5.10(b) Machine-generated source records shall identify the system, model, device, node, software, API, dashboard, or infrastructure that generated the source; owner where known; operator where applicable; provider where any; version; configuration; input source; processing method; output type; time; custody; confidence where meaningful; uncertainty; limitations; public-safe status; and correction path.
7.5.10(c) Machine-generated sources shall not be treated as neutral, complete, objective, authoritative, current, accurate, public-safe, or decision-grade merely because they are automated, mathematical, cryptographic, real-time, model-based, distributed, signed, anchored, or visually precise.
7.5.10(d) AI-generated source treatment shall include model identity where material, prompt or input category, retrieval source, training or inference restriction, human review where material, hallucination risk, bias risk, drift risk, source-grounding status, public-safe status, and correction path.
7.5.10(e) Sensor, telemetry, DePIN, and AI-RAN source treatment shall include device or node identity, calibration or signal quality, timing, location or safe location treatment, spoof risk, tamper risk, missing data, stale data, processing method, confidence, uncertainty, and correction path.
7.5.10(f) Blockchain, proof receipt, smart contract, hash, anchor, role-key, or entitlement-state source treatment shall include source record relationship, signer or key authority, custody, timestamp, correction layer, revocation or supersession path, and limitation that technical proof does not itself create institutional authority.
7.5.10(g) Digital twin, simulation, scenario, and model-based source treatment shall include assumptions, input records, calibration, validation, sensitivity, uncertainty, limitations, and public-safe status.
7.5.10(h) Machine-generated sources shall not create public authority decisions, official warnings, emergency commands, certification, recognition, finance-readiness, procurement approval, provider endorsement, ratings, guarantees, operational commands, protocol effect, legal status, or execution consequence by default.
7.5.10(i) The controlling rule shall be that machine generation changes the source treatment required; it does not reduce the need for source treatment.
7.5.11 Derived Evidence and Downstream Dependency Tracking. 7.5.11(a) Derived evidence shall include evidence created through processing, aggregation, fusion, summarization, translation, redaction, anonymization, de-identification, mapping, dashboarding, modeling, AI-assisted interpretation, statistical analysis, source comparison, benchmark analysis, public-safe transformation, controlled annex preparation, Docket input creation, Grid input creation, Observatory record creation, Truth Engine comparison, Rails handoff, GRF input, GRA input, Protocol Authority support, or publication.
7.5.11(b) Derived evidence records shall identify all material upstream sources or source classes, transformation methods, processing steps, assumptions, exclusions, confidence effects, uncertainty effects, limitation effects, sensitivity effects, public-safe effects, and correction dependencies.
7.5.11(c) Derived evidence shall preserve source restrictions, permissions, licenses, consent or non-consent treatment where applicable, public authority restrictions, privacy restrictions, cybersecurity restrictions, sovereign data restrictions, community safeguards, protected knowledge controls, IP restrictions, finance-facing limits, procurement-facing limits, provider-neutrality limits, sponsor non-control limits, and publication limits.
7.5.11(d) Derivation shall not erase uncertainty, conflict, missing data, source bias, source conflict, source sensitivity, public authority context, community context, protected knowledge status, or correction obligations.
7.5.11(e) Downstream dependency tracking shall identify which reports, dashboards, maps, models, Academy materials, Docket records, Grid records, Observatory records, Truth Engine records, Rails handoffs, GRF inputs, GRA inputs, Protocol Authority support, public authority materials, finance-facing materials, procurement-facing materials, provider materials, sponsor materials, National Company materials, Project SPV materials, public-good software, technical baselines, public claims, and archives depend on a source or derived evidence record.
7.5.11(f) Where an upstream source is corrected, superseded, withdrawn, downgraded, restricted, reclassified, disputed, invalidated, or found unsafe, GCRI Canada shall review downstream dependencies and determine whether correction, supersession, withdrawal, retraction, downgrade, restriction, notification, or public-safe clarification is required.
7.5.11(g) Derived evidence shall not create authority beyond the source records. Aggregation, fusion, AI analysis, dashboard presentation, or multi-source corroboration shall not create public authority effect, certification, recognition, finance-readiness, procurement approval, provider endorsement, sponsor control, rating, guarantee, public warning, emergency command, or execution consequence by default.
7.5.11(h) The controlling rule shall be that derived evidence inherits the responsibilities of its sources and creates new responsibilities for every downstream use.
7.5.12 Source Lineage Correction, Supersession, Withdrawal, and Archival. 7.5.12(a) Source lineage, provenance, and custody records shall be subject to correction, supersession, withdrawal, restriction, reclassification, downgrade, reinstatement, archival annotation, downstream dependency review, and closeout.
7.5.12(b) Correction shall be required where source lineage is inaccurate, incomplete, misattributed, circular, broken, misleading, stale, unauthorized, improperly permissioned, wrongly classified, missing conflicts, missing public authority capacity, missing community safeguards, missing provenance, missing custody, or inconsistent with source records.
7.5.12(c) Supersession shall be used where a later source record, corrected provenance record, corrected custody record, corrected permission record, corrected public authority record, corrected community safeguard record, corrected machine-generated record, or corrected derived evidence record replaces an earlier record while preserving institutional memory.
7.5.12(d) Withdrawal shall be used where the source cannot be lawfully, safely, or reliably used; where permission is absent or revoked; where public authority authority is missing; where protected knowledge was mishandled; where cyber-sensitive information is unsafe; where privacy or sovereign data restrictions prohibit use; or where source deficiencies cannot be corrected in place.
7.5.12(e) Restriction or reclassification shall be used where source lineage remains valid but sensitivity, public-safe status, public authority status, finance-facing status, procurement-facing status, provider-facing status, sponsor-facing status, community status, cyber status, infrastructure status, or legal status requires a more protective handling category.
7.5.12(f) Archival shall preserve prior source lineage, provenance, custody, correction history, supersession status, withdrawal status, restriction status, reclassification status, access limits, retention basis, and future retrieval limits.
7.5.12(g) Where source lineage correction affects public materials, public authority materials, finance-facing materials, procurement-facing materials, provider materials, sponsor materials, community-facing materials, Academy materials, Docket records, Grid records, Observatory records, Truth Engine records, Rails handoffs, GRF inputs, GRA inputs, Protocol Authority support, National Company materials, Project SPV materials, dashboards, maps, repositories, APIs, technical baselines, or public claims, GCRI Canada shall conduct downstream dependency review and correct affected materials where within its control or interface rights.
7.5.12(h) The controlling rule shall be that source lineage must itself be correctable because evidence integrity depends on the integrity of the path from source to use.
7.6 Evidence Classification and Handling
7.6.1 Public Evidence. 7.6.1(a) Public Evidence shall mean evidence that GCRI Canada has determined may be disclosed publicly without unreasonable risk to privacy, cybersecurity, sovereign data, public authority confidentiality, protected knowledge, community safety, infrastructure security, commercial confidentiality, finance-boundary discipline, procurement neutrality, provider neutrality, sponsor non-control, public-safe communication, or public trust.
7.6.1(b) Public Evidence may include public reports, public datasets, public-safe summaries, public technical notes, public-good software documentation, released technical baseline summaries, public Academy materials, public Docket summaries, public Grid summaries, public Observatory summaries, public correction notices, supersession notices, withdrawal notices, retraction notices, and public-facing explanatory materials.
7.6.1(c) Public classification shall require confirmation that the evidence source, license, permission, data-sharing terms, public authority restrictions, community safeguards, protected knowledge restrictions, IP rights, cybersecurity risks, infrastructure sensitivity, finance sensitivity, procurement sensitivity, commercial sensitivity, and public-safe limits permit public disclosure.
7.6.1(d) Public Evidence shall remain subject to source lineage, provenance, versioning, limitations, confidence and uncertainty treatment, public claims discipline, correctionability, supersession, withdrawal, retraction, and archival controls.
7.6.1(e) Public classification shall not imply certification, recognition, finance-readiness, public authority approval, procurement approval, provider endorsement, sponsor approval, official warning, emergency command, operational clearance, legal advice, professional advice, rating, guarantee, or execution consequence.
7.6.1(f) Public Evidence shall be published with appropriate context, limitations, public-safe framing, non-authority language, non-warning language, non-finance-readiness language where material, non-procurement language where material, non-provider-endorsement language where material, and correction path.
7.6.1(g) Public Evidence may be reclassified, restricted, corrected, superseded, withdrawn, or retracted where later review identifies error, stale status, unsafe disclosure, changed legal context, changed public authority context, changed public-safe risk, rights risk, community concern, cybersecurity concern, protected knowledge concern, or material public misunderstanding.
7.6.1(h) The controlling rule shall be that Public Evidence may be open to the public, but it remains bounded, source-lined, non-executing, and correctable.
7.6.2 Public-Safe Evidence. 7.6.2(a) Public-Safe Evidence shall mean evidence that may be communicated, summarized, visualized, taught, published, or referenced in a form designed to protect against avoidable harm, false authority, public warning confusion, emergency command confusion, finance overclaim, procurement implication, provider preference, sponsor control, privacy harm, cybersecurity harm, community harm, protected knowledge exposure, infrastructure exposure, reputational harm, market distortion, or public trust erosion.
7.6.2(b) Public-Safe Evidence may include evidence that is safe for public release only after aggregation, generalization, redaction, masking, delayed publication, controlled vocabulary treatment, limitation language, visual redesign, confidence and uncertainty disclosure, source protection, or removal of sensitive details.
7.6.2(c) Public-Safe classification shall require review of audience, purpose, source lineage, evidence quality, data class, public authority context, community context, infrastructure context, cyber context, health context, finance-facing context, procurement-facing context, provider context, sponsor context, media risk, and public reliance risk.
7.6.2(d) Public-Safe Evidence shall not necessarily mean fully public evidence. Some evidence may be public-safe only for a defined audience, defined room, defined public authority learning context, controlled annex, Academy context, Docket context, Grid context, Observatory summary, Rails handoff, or interface record.
7.6.2(e) Public-Safe Evidence shall include boundary language where material, including no public authority action, no public warning, no emergency command, no certification, no recognition, no finance-readiness, no investment advice, no insurance, no underwriting, no rating, no public finance approval, no procurement preference, no provider endorsement, no sponsor control, no professional advice, and no execution.
7.6.2(f) Maps, dashboards, charts, scores, indicators, colors, icons, badges, labels, rankings, live-status displays, and AI-generated summaries shall be reviewed as public claims instruments and shall not visually imply authority, urgency, certainty, approval, ranking, warning, command, finance-readiness, certification, recognition, procurement status, or provider preference beyond the record.
7.6.2(g) Public-Safe Evidence shall remain correctable. Where public-safe framing later proves misleading, incomplete, unsafe, stale, overbroad, under-protective, or authority-inflating, GCRI Canada shall correct, restrict, supersede, withdraw, retract, or reclassify affected materials.
7.6.2(h) The controlling rule shall be that Public-Safe Evidence is evidence communicated with safeguards, not evidence freed from safeguards.
7.6.3 Internal Evidence. 7.6.3(a) Internal Evidence shall mean evidence available for internal GCRI Canada use within authorized roles, systems, repositories, rooms, workflows, committees, research programs, technical stewardship processes, public authority learning preparation, public-safe publication review, Docket preparation, Grid preparation, Observatory method development, Truth Engine method development, Rails preparation, Academy development, interface preparation, or correction processes.
7.6.3(b) Internal classification shall be used where evidence is not approved for public release, external handoff, public authority circulation, finance-facing use, provider-facing use, sponsor-facing use, community-facing use, or public claims, but may be used internally for research, evidence review, method development, classification, safeguard design, technical baseline development, public-good software development, or correction.
7.6.3(c) Internal Evidence shall be accessed only by persons with a legitimate institutional need, appropriate role, confidentiality obligation where applicable, data handling authority, cybersecurity clearance where applicable, and training appropriate to the evidence class.
7.6.3(d) Internal classification shall not be used to conceal error, avoid correction, suppress dissent, avoid public-safe clarification, protect sponsor or provider interests, hide conflicts, or evade public-benefit accountability where lawful and safe clarification is required.
7.6.3(e) Internal Evidence may include draft methods, draft reports, preliminary datasets, early observations, internal review notes, internal challenge records, method-development materials, issue logs, correction investigations, technical review notes, and materials awaiting classification.
7.6.3(f) Internal Evidence shall not be cited externally, described as GCRI Canada finding, used for public claims, routed to public authorities, routed to finance-facing actors, used in procurement-facing materials, or relied upon for external status unless reclassified through appropriate review.
7.6.3(g) Where Internal Evidence becomes public, external, public authority-facing, finance-facing, provider-facing, sponsor-facing, community-facing, or public-safe, the evidence shall be re-reviewed and reclassified before such use.
7.6.3(h) The controlling rule shall be that Internal Evidence supports institutional work inside the boundary, not public meaning outside the boundary.
7.6.4 Confidential Evidence. 7.6.4(a) Confidential Evidence shall mean evidence subject to confidentiality obligations, non-disclosure duties, contractual restrictions, public authority confidentiality, research confidentiality, source protection, commercial confidentiality, personal information protection, community restrictions, protected knowledge controls, provider restrictions, sponsor restrictions, or other lawful confidentiality limits.
7.6.4(b) Confidential classification shall identify the source of confidentiality, scope, duration, permitted recipients, prohibited recipients, permitted uses, prohibited uses, publication limits, onward-sharing limits, retention obligations, destruction or return obligations where applicable, breach procedures, and correction path.
7.6.4(c) Confidential Evidence shall be stored, accessed, transmitted, processed, summarized, modeled, dashboarded, mapped, routed, and archived only through authorized systems and personnel consistent with confidentiality obligations and cybersecurity controls.
7.6.4(d) Confidential Evidence shall not be disclosed publicly, placed in public repositories, used in public dashboards, used in public maps, included in public Academy materials, included in public-safe reports, disclosed to sponsors, disclosed to providers, disclosed to capital readers, disclosed to public authorities, or disclosed to communities unless the confidentiality terms permit such disclosure and public-safe review approves it.
7.6.4(e) Confidentiality shall not convert evidence into sponsor-controlled evidence, provider-controlled evidence, public authority-controlled evidence, or private truth. GCRI Canada shall preserve independence, conflict review, public-good purpose, and correctionability even where confidentiality applies.
7.6.4(f) Where a correction is required for Confidential Evidence, GCRI Canada shall correct within confidentiality limits and shall use controlled correction notices, restricted notices, anonymized notices, public-safe clarifications, or legal review where needed to balance confidentiality and correctionability.
7.6.4(g) Where Confidential Evidence is improperly disclosed, misused, overclaimed, or cited beyond permitted scope, GCRI Canada shall restrict further access, investigate, notify affected actors where appropriate and lawful, correct derivative materials, and preserve incident and correction records.
7.6.4(h) The controlling rule shall be that confidentiality controls disclosure, not truth, correction, or institutional independence.
7.6.5 Restricted Evidence. 7.6.5(a) Restricted Evidence shall mean evidence requiring heightened access limitation because of legal, privacy, cybersecurity, public authority, health, infrastructure, community, protected knowledge, national security-adjacent, export-control, sanctions, controlled technology, commercial, finance, procurement, source-protection, or public-safe risk.
7.6.5(b) Restricted classification shall identify restriction basis, authorized roles, access conditions, system controls, logging requirements, storage requirements, transfer limits, AI-use limits, publication prohibition or limits, handoff restrictions, retention limits, and correction path.
7.6.5(c) Restricted Evidence may include cyber vulnerabilities, incident records, public authority sensitive records, health-sensitive data, infrastructure-sensitive data, detailed geospatial layers, community-protected knowledge, identity data, finance-facing non-public records, procurement-sensitive materials, controlled technology information, export-control-sensitive information, sanctions-sensitive information, or source-protected records.
7.6.5(d) Restricted Evidence shall not be copied, exported, summarized, translated, embedded, trained on, modeled, dashboarded, mapped, published, or routed outside authorized controls unless reclassification or a permitted controlled handoff is recorded.
7.6.5(e) Restricted Evidence shall require need-to-know access, access logs where appropriate, periodic access review, incident procedures, and downstream dependency tracking.
7.6.5(f) Restricted classification shall not create secrecy for convenience, sponsor control, provider control, public authority overreach, or avoidance of correction. Restriction shall be proportionate to risk and shall remain subject to review.
7.6.5(g) Where Restricted Evidence is required for public-safe understanding, GCRI Canada may create public-safe summaries, controlled summaries, aggregated summaries, redacted summaries, or no-disclosure records that preserve public benefit without exposing restricted content.
7.6.5(h) The controlling rule shall be that Restricted Evidence may be institutionally important precisely because it cannot be freely disclosed.
7.6.6 Public Authority Evidence. 7.6.6(a) Public Authority Evidence shall mean evidence received from, concerning, contributed by, generated by, reviewed by, or materially involving a public authority, public official, public body, Crown entity, municipality, Indigenous government or authority, provincial or territorial authority, federal body, regulator, emergency management actor, public health actor, public safety actor, public works actor, public finance actor, procurement actor, or other public actor.
7.6.6(b) Public Authority Evidence shall be recorded with source, capacity classification, authority to contribute where applicable, official classification, permitted use, prohibited use, publication permission, reference approval, confidentiality, retention, transfer limits, correction rights, withdrawal or restriction path, and non-public-authority-action status.
7.6.6(c) Public Authority Evidence shall not create public authority delegation, endorsement, adoption, official guidance, regulatory approval, public service decision, public warning, emergency command, procurement approval, funding approval, public finance approval, public-private partnership, sovereign obligation, or public authority action by GCRI Canada.
7.6.6(d) Public authority names, logos, titles, quotes, photographs, agency names, jurisdiction names, attendance references, venue references, data references, and funding references shall be handled under reference controls to prevent apparent authority.
7.6.6(e) Public Authority Evidence may be used for public authority learning, evidence literacy, technical literacy, public-safe reports, controlled annexes, Academy materials, Docket inputs, Grid inputs, Observatory methods, Truth Engine methods, Rails handoffs, GRF inputs, GRA inputs, Protocol Authority support, and correction records only within the recorded capacity and permitted use.
7.6.6(f) Where Public Authority Evidence is confidential, sensitive, emergency-adjacent, public health-sensitive, public safety-sensitive, procurement-sensitive, public finance-sensitive, infrastructure-sensitive, or rights-bearing, additional classification controls shall apply.
7.6.6(g) Where Public Authority Evidence is misdescribed, overclaimed, misattributed, corrected, restricted, withdrawn, superseded, or reclassified, GCRI Canada shall correct affected materials and notify or coordinate with the relevant public authority where appropriate and lawful.
7.6.6(h) The controlling rule shall be that public authority evidence must preserve the line between public authority context and public authority act.
7.6.7 Health-Sensitive Evidence. 7.6.7(a) Health-Sensitive Evidence shall mean evidence involving health data, public health data, clinical context, biosecurity context, health-adjacent sensors, wastewater signals, vulnerable persons, community health data, public health signals, biosecurity evidence, health system evidence, health-sensitive geospatial data, or other information whose use or disclosure may affect health privacy, public health interpretation, clinical reliance, public authority duties, stigma, discrimination, or public trust.
7.6.7(b) Health-Sensitive Evidence shall be recorded with source, lawful basis or authority to contribute where applicable, data class, identifiability, consent or non-consent treatment where applicable, ethics review where applicable, public authority context where any, health authority context where any, permitted use, prohibited use, publication limits, AI-use limits, privacy safeguards, public-safe status, and correction path.
7.6.7(c) Health-Sensitive Evidence shall not create clinical advice, diagnosis, treatment advice, public health order, public warning, emergency command, public authority decision, regulatory approval, certification, insurance rating, finance-readiness, procurement approval, provider endorsement, or execution consequence by default.
7.6.7(d) Health-Sensitive Evidence shall be handled with heightened minimization, access control, de-identification where appropriate and sufficient, aggregation where appropriate, public-safe publication review, stigma review, re-identification review, small-population review, AI-use review, and correction controls.
7.6.7(e) Health-sensitive dashboards, maps, reports, and models shall avoid false precision, clinical implication, public health order implication, public warning implication, stigma, group harm, community harm, and unsafe reliance.
7.6.7(f) Biosecurity-sensitive evidence shall receive additional controlled-access, misuse-prevention, public-safe publication, and legal review where material.
7.6.7(g) Where Health-Sensitive Evidence is misused, over-disclosed, misclassified, corrected, superseded, withdrawn, restricted, or reclassified, GCRI Canada shall review downstream materials and preserve correction and safeguard records.
7.6.7(h) The controlling rule shall be that health-sensitive evidence must support learning without becoming clinical authority, public health authority, or public harm.
7.6.8 Cyber-Sensitive Evidence. 7.6.8(a) Cyber-Sensitive Evidence shall mean evidence involving cyber vulnerabilities, exploits, incident records, threat indicators, access logs, network logs, endpoint logs, repository logs, build pipeline records, secrets, keys, authentication data, operational technology, industrial control systems, cyber-physical systems, public authority systems, critical infrastructure, software supply chains, or security findings whose disclosure or misuse may increase harm.
7.6.8(b) Cyber-Sensitive Evidence shall be recorded with source, asset or system context, owner where known, operator where applicable, custodian, vulnerability or incident status where applicable, sensitivity level, collection method, custody, integrity controls, disclosure restrictions, permitted use, prohibited use, access controls, public-safe status, and correction path.
7.6.8(c) Cyber-Sensitive Evidence shall not create law enforcement authority, regulatory authority, public authority finding, official warning, emergency command, managed security provider status, certification, insurance rating, finance-readiness, procurement approval, provider endorsement, operational command, or execution consequence by default.
7.6.8(d) Cyber-Sensitive Evidence shall be handled through need-to-know access, secure storage, secure transmission, controlled rooms where needed, secure disclosure procedures, redaction where needed, vulnerability disclosure controls, public-safe publication review, legal review where material, and incident procedures.
7.6.8(e) Public-safe summaries of Cyber-Sensitive Evidence shall avoid exploit-enabling detail, attack-path disclosure, credential exposure, critical dependency exposure, source exposure, public authority system exposure, infrastructure weakness exposure, and unsafe technical specificity.
7.6.8(f) Repository security, dependency security, SBOM-related records where applicable, key management, secret handling, vulnerability management, and release security shall be considered in handling Cyber-Sensitive Evidence connected to public-good software or technical baselines.
7.6.8(g) Where Cyber-Sensitive Evidence is improperly disclosed, corrected, superseded, patched, mitigated, exploited, reclassified, withdrawn, or found unsafe, GCRI Canada shall review affected materials, public-safe reports, technical baselines, repositories, dashboards, Docket records, Grid records, Observatory records, Truth Engine records, Rails handoffs, and public claims.
7.6.8(h) The controlling rule shall be that cyber-sensitive evidence must reduce risk without becoming a risk.
7.6.9 Infrastructure-Sensitive Evidence. 7.6.9(a) Infrastructure-Sensitive Evidence shall mean evidence concerning critical infrastructure, energy systems, grids, utilities, water systems, telecommunications systems, ports, corridors, transportation systems, industrial systems, operational technology, public authority facilities, emergency systems, remote connectivity systems, compute systems, supply chains, logistics systems, health-adjacent facilities, and mission-critical infrastructure where disclosure or misuse could create operational, public safety, security, market, community, or public trust risk.
7.6.9(b) Infrastructure-Sensitive Evidence shall be recorded with source, asset or system context, owner where known, operator where applicable, host where applicable, provider where any, data class, operational sensitivity, cyber-physical sensitivity, geospatial sensitivity, public authority relevance, public-safe status, permitted use, prohibited use, access controls, and correction path.
7.6.9(c) Infrastructure-Sensitive Evidence shall not create infrastructure operation authority, utility direction, engineering certification, public warning, emergency command, public authority decision, procurement approval, finance-readiness, insurance rating, credit rating, public finance approval, provider endorsement, guarantee, or execution consequence by default.
7.6.9(d) Infrastructure-Sensitive Evidence may include asset locations, topology, dependency maps, outage patterns, repair constraints, fuel dependencies, storage capacity, backup power status, telemetry, vulnerabilities, access points, security controls, operational constraints, critical loads, chokepoints, spare parts dependencies, supplier dependencies, and facility capabilities.
7.6.9(e) Public-safe handling shall use aggregation, generalization, redaction, masking, delayed release, controlled annexes, restricted access, or refusal where full disclosure could expose weakness, attack surface, sabotage opportunity, market manipulation, community vulnerability, or public panic risk.
7.6.9(f) Maps, dashboards, and visualizations using Infrastructure-Sensitive Evidence shall be reviewed for location precision, dependency exposure, public warning implication, emergency command implication, provider implication, procurement implication, finance implication, and public-safe status.
7.6.9(g) Where Infrastructure-Sensitive Evidence is over-disclosed, corrected, superseded, restricted, reclassified, withdrawn, or found unsafe, GCRI Canada shall restrict access where possible, correct derivative materials, notify affected actors where appropriate and lawful, and preserve incident and correction records.
7.6.9(h) The controlling rule shall be that infrastructure-sensitive evidence must strengthen resilience without revealing how to disrupt it or falsely implying who controls it.
7.6.10 Finance-Sensitive Evidence. 7.6.10(a) Finance-Sensitive Evidence shall mean evidence that may be used, misused, or misunderstood in finance-facing, capital-reader, insurance, underwriting, lending, rating, public finance, investment, securities, guarantee, National Company, Project SPV, procurement, or transaction contexts.
7.6.10(b) Finance-Sensitive Evidence shall be recorded with source, domain, intended finance-facing context where any, public-safe status, finance-safe status, confidence, uncertainty, limitations, no-investment-advice status, no-offer status, no-solicitation status, no-brokerage status, no-finder status, no-placement status, no-lending status, no-insurance status, no-underwriting status, no-rating status, no-public-finance-approval status, no-guarantee status, permitted use, prohibited use, and correction path.
7.6.10(c) Finance-Sensitive Evidence shall not create finance-readiness, investment advice, securities offering, brokerage, finder activity, placement, lending, credit approval, insurance approval, underwriting conclusion, rating, public finance approval, guarantee, capital commitment, transaction approval, fund approval, market operation, procurement approval, or GCRI Canada finance authority.
7.6.10(d) Finance-Sensitive Evidence may be routed to GRA, Nexus Rails, public finance readers, capital readers, insurers, lenders, underwriters, rating actors, National Companies, Project SPVs, public authorities, or other competent actors only through boundary-safe handoff records and regulated-perimeter controls where material.
7.6.10(e) Scores, indicators, maturity-context records, resilience-context records, readiness-context records, dashboards, maps, benchmarks, and evidence packs shall not be described as ratings, guarantees, underwriting conclusions, insurance signals, credit signals, investment signals, public finance determinations, procurement rankings, or capital-readiness determinations by GCRI Canada.
7.6.10(f) Where finance-facing use is ambiguous, GCRI Canada shall apply regulated-perimeter review, stronger boundary language, access restriction, GRA routing, legal review, or refusal.
7.6.10(g) Where Finance-Sensitive Evidence is misused, overclaimed, corrected, superseded, withdrawn, restricted, or reclassified, GCRI Canada shall review affected finance-facing materials, public claims, Rails handoffs, GRA inputs, National Company materials, Project SPV materials, capital-reader materials, and public-safe reports.
7.6.10(h) The controlling rule shall be that finance-sensitive evidence must be handled so that evidence support is never mistaken for financial execution.
7.6.11 Commercially Sensitive Evidence. 7.6.11(a) Commercially Sensitive Evidence shall mean evidence involving confidential business information, trade secrets, proprietary methods, supplier identity, pricing, capacity, customer information, production volumes, lead times, inventory, facility capabilities, process know-how, product roadmaps, provider configurations, contractual restrictions, commercial relationships, competitive strategy, or market-sensitive information.
7.6.11(b) Commercially Sensitive Evidence shall be recorded with source, owner where known, confidentiality basis, IP status, license or use terms, commercial sensitivity, permitted use, prohibited use, access controls, publication limits, provider-neutrality status, competition-safety status, and correction path.
7.6.11(c) Commercially Sensitive Evidence shall not create provider endorsement, supplier ranking, market allocation, procurement preference, certification, recognition, finance-readiness, public authority approval, market superiority, rating, guarantee, operational clearance, or execution consequence by default.
7.6.11(d) Commercially Sensitive Evidence shall be reviewed for competition risk, procurement steering risk, provider-preference risk, sponsor influence risk, unfair advantage risk, exclusionary effect, market distortion, confidentiality obligations, and IP restrictions.
7.6.11(e) GCRI Canada shall not disclose Commercially Sensitive Evidence to other providers, sponsors, competitors, public authorities, capital readers, National Companies, Project SPVs, media, or the public unless disclosure is permitted, lawful, public-safe, competition-safe, and consistent with the source record.
7.6.11(f) Public-safe summaries may describe general method lessons, evidence categories, resilience patterns, safeguard lessons, and correction needs without disclosing proprietary details, trade secrets, supplier dependencies, commercially sensitive comparisons, or unfair competitive information.
7.6.11(g) Where Commercially Sensitive Evidence is misused, over-disclosed, misclassified, corrected, superseded, withdrawn, restricted, or reclassified, GCRI Canada shall review downstream materials, notify affected actors where appropriate and lawful, and preserve incident and correction records.
7.6.11(h) The controlling rule shall be that commercially sensitive evidence may support public-good learning only where commercial sensitivity, competition safety, and public-good independence are preserved.
7.6.12 Personal Information and Rights-Bearing Evidence. 7.6.12(a) Personal Information and Rights-Bearing Evidence shall mean evidence involving identifiable persons, reasonably identifiable persons, households, groups, communities, identity data, location data, mobility data, health data, financial data, public service data, employment-adjacent data, education-adjacent data, biometric or biometric-adjacent data, device data, consent records, access records, protected participation, vulnerability indicators, or inferred status information.
7.6.12(b) Personal Information and Rights-Bearing Evidence shall be recorded with source, lawful basis where applicable, consent or non-consent treatment where applicable, data subject or affected group context where appropriate, data class, sensitivity, purpose, minimization treatment, access controls, retention, transfer limits, publication limits, AI-use limits, correction rights, withdrawal or restriction path, and public-safe status.
7.6.12(c) Such evidence shall not create public authority decisions, eligibility determinations, entitlement determinations, clinical advice, public health orders, financial decisions, insurance decisions, credit decisions, employment decisions, procurement decisions, certification, recognition, finance-readiness, provider endorsement, public warning, emergency command, or execution consequence by GCRI Canada.
7.6.12(d) GCRI Canada shall apply minimization, role-based access, de-identification where appropriate and sufficient, aggregation where appropriate, compute-to-data where needed, AI-use restriction, publication restriction, security controls, and public-safe review.
7.6.12(e) Rights-bearing evidence shall be reviewed for inclusion, exclusion, discrimination, accessibility, re-identification, inference, surveillance, stigma, retaliation, source exposure, community harm, and public authority misuse.
7.6.12(f) Public-facing materials shall not identify persons or affected groups in ways that create avoidable harm, stigma, unsafe reliance, discrimination, retaliation, public authority misuse, or loss of dignity.
7.6.12(g) Where Personal Information or Rights-Bearing Evidence is misused, over-disclosed, incorrectly classified, corrected, withdrawn, restricted, or reclassified, GCRI Canada shall restrict access, correct derivative materials, notify affected actors where appropriate and lawful, review downstream dependencies, and preserve incident and correction records.
7.6.12(h) The controlling rule shall be that rights-bearing evidence must preserve the rights, dignity, safety, and correction interests of persons and communities before it serves institutional analysis.
7.6.13 Community-Protected Evidence. 7.6.13(a) Community-Protected Evidence shall mean evidence contributed by, concerning, or materially affecting communities where use or disclosure may affect community safety, dignity, autonomy, reputation, service access, public authority treatment, protected participation, location privacy, cultural integrity, economic opportunity, environmental protection, or public trust.
7.6.13(b) Community-Protected Evidence shall be recorded with community context, contributor authority, source, permitted use, prohibited use, consent or non-consent treatment where applicable, attribution or non-attribution, community protocol where applicable, publication limits, mapping limits, AI-use limits, data-sharing limits, grievance path, withdrawal or restriction path, and correction path.
7.6.13(c) Community-Protected Evidence shall not create community consent, community endorsement, public authority mandate, public warning, certification, recognition, finance-readiness, procurement approval, provider endorsement, sponsor approval, media permission, public-facing legitimacy, or execution consequence beyond the recorded scope.
7.6.13(d) Public-safe handling shall account for small-population identifiability, sensitive locations, community vulnerability, service gaps, public authority relationships, historical under-service, stigma, surveillance risk, retaliation risk, media misuse, sponsor misuse, provider misuse, finance misuse, and public trust harm.
7.6.13(e) GCRI Canada shall not convert Community-Protected Evidence into generic public data, sponsor impact narrative, provider marketing, public-facing map, AI training material, finance-facing claim, or media story without appropriate permission, safeguard review, and public-safe treatment.
7.6.13(f) Where community-protected material must be communicated, GCRI Canada shall use aggregation, generalization, masking, redaction, local review where appropriate, controlled annexes, restricted handoffs, or public-safe summaries.
7.6.13(g) Where Community-Protected Evidence is misused, over-disclosed, misattributed, overclaimed, incorrectly mapped, corrected, restricted, withdrawn, or reclassified, GCRI Canada shall review downstream dependencies and notify affected community actors where appropriate and lawful.
7.6.13(h) The controlling rule shall be that community-protected evidence is not less rigorous because it is contextual; it is more demanding because harm can travel through context.
7.6.14 Indigenous, Local, Territorial, Cultural, Environmental, and Protected Knowledge Evidence. 7.6.14(a) Indigenous, local, territorial, cultural, environmental, and Protected Knowledge Evidence shall mean evidence grounded in Indigenous knowledge, local knowledge, territorial context, cultural practice, land-based knowledge, water-based knowledge, food-system knowledge, ecological knowledge, environmental observation, sacred or culturally sensitive information, protected site information, knowledge-holder contribution, or community-governed knowledge.
7.6.14(b) Such evidence shall be recorded with contributor authority, knowledge-holder context, community or territorial context, Indigenous data sovereignty where applicable, community protocol, cultural protocol, consent or non-consent treatment where applicable, attribution or non-attribution, permitted use, prohibited use, publication limits, public-safe mapping limits, AI-use limits, data-sharing limits, grievance path, withdrawal or restriction path, and correction path.
7.6.14(c) Protected Knowledge Evidence shall not be treated as ordinary open data, generic public evidence, sponsor-reporting content, provider-testing content, public map content, AI-training content, media content, finance-facing material, or public-facing narrative merely because it has institutional value.
7.6.14(d) GCRI Canada shall protect sacred sites, cultural sites, protected habitats, sensitive ecological locations, food sources, water sources, routes, households, knowledge holders, ceremonial information, local hazard knowledge, and culturally governed information from unsafe disclosure, extraction, decontextualization, or enclosure.
7.6.14(e) Indigenous or local knowledge contribution shall not create consent, authorization, endorsement, public authority mandate, public authority approval, recognition, certification, finance-readiness, procurement approval, provider endorsement, sponsor control, public warning, emergency command, or execution consequence beyond the recorded scope.
7.6.14(f) Where public-safe communication is possible, it shall preserve context, permissions, limitations, attribution or non-attribution, protected knowledge limits, and correction rights.
7.6.14(g) Where protected knowledge is misused, over-disclosed, misattributed, incorrectly mapped, extracted, enclosed, corrected, restricted, withdrawn, or reclassified, GCRI Canada shall take protective correction steps, review downstream dependencies, and notify affected knowledge holders where appropriate and lawful.
7.6.14(h) The controlling rule shall be that protected knowledge must remain protected even when it supports public-good evidence.
7.6.15 Controlled Technology and Export-Control-Sensitive Evidence. 7.6.15(a) Controlled Technology and Export-Control-Sensitive Evidence shall mean evidence involving controlled technology, dual-use technology, export-controlled information, sanctions-sensitive information, cryptographic systems, cyber tools, satellite systems, telecommunications systems, AI systems, semiconductor systems, advanced manufacturing processes, materials data, operational technology, national security-adjacent systems, restricted software, restricted equipment, technical data, or know-how whose handling may be subject to legal, security, trade, or public-safe restrictions.
7.6.15(b) Such evidence shall be recorded with technology class, source, owner where known, jurisdiction, recipient restrictions, transfer restrictions, publication restrictions, repository restrictions, cloud or compute restrictions, export-control review status, sanctions review status, controlled-technology status, dual-use risk, access controls, permitted use, prohibited use, and correction path.
7.6.15(c) Controlled Technology and Export-Control-Sensitive Evidence shall not create trade authority, export-control approval, sanctions clearance, certification, procurement approval, provider endorsement, finance-readiness, public authority approval, operational clearance, market authority, or execution consequence by default.
7.6.15(d) GCRI Canada shall restrict access, storage, transmission, publication, repository placement, model training, embedding, cross-border transfer, public-good software release, technical baseline release, dashboarding, mapping, and external handoff where required by law, security, contract, or public-safe review.
7.6.15(e) Where export-control, sanctions, controlled-technology, dual-use, national security-adjacent, or trade-restriction uncertainty exists, GCRI Canada shall hold, restrict, re-scope, localize, compute-to-data, seek legal review, route to competent actors, refuse participation, or withdraw materials until the risk is resolved.
7.6.15(f) Public-safe summaries shall not disclose restricted technical details, controlled know-how, exploit-enabling information, sensitive specifications, sensitive configurations, restricted transfer information, or details that could facilitate misuse.
7.6.15(g) Where Controlled Technology or Export-Control-Sensitive Evidence is misclassified, over-disclosed, transferred improperly, corrected, withdrawn, restricted, or reclassified, GCRI Canada shall investigate, restrict further use, seek legal review where material, notify affected actors where appropriate and lawful, and preserve incident and correction records.
7.6.15(h) The controlling rule shall be that evidence concerning controlled technology must not become an uncontrolled transfer.
7.6.16 Classification, Access, Retention, Redaction, Publication, and Correction Rules by Evidence Class. 7.6.16(a) GCRI Canada shall maintain classification, access, retention, redaction, publication, and correction rules by evidence class, and shall apply the most protective applicable rule where evidence falls within more than one class or where classification ambiguity exists.
7.6.16(b) Classification rules shall identify evidence class, sensitivity, public-safe status, public authority status, privacy status, cybersecurity status, sovereignty status, community or protected knowledge status, finance-facing status, procurement-facing status, provider-facing status, sponsor-facing status, commercial status, controlled-technology status, and permitted use.
7.6.16(c) Access rules shall identify authorized roles, need-to-know requirements, training requirements, confidentiality obligations, system permissions, controlled-room requirements, clean-room requirements, logging requirements, access review cycles, and suspension or revocation conditions.
7.6.16(d) Retention rules shall identify retention period, retention basis, review cycle, archival status, deletion or restriction conditions, legal hold where applicable, public authority restrictions, data-sharing obligations, community restrictions, protected knowledge restrictions, and correction obligations.
7.6.16(e) Redaction rules shall identify what may be removed, masked, generalized, aggregated, delayed, de-identified where appropriate and sufficient, summarized, or withheld before publication, handoff, dashboarding, mapping, Academy use, Docket use, Grid use, Observatory use, Truth Engine use, Rails use, GRF input, GRA input, Protocol Authority support, or archive access.
7.6.16(f) Publication rules shall identify whether evidence may be public, public-safe, internal, controlled, restricted, confidential, public authority-facing, finance-facing, provider-facing, sponsor-facing, community-facing, or withheld, and shall require public-safe review before public disclosure.
7.6.16(g) Correction rules shall identify who may correct, who must be notified where appropriate, what derivative materials require review, whether correction may be public or controlled, whether withdrawal or retraction is required, whether archive annotation is required, and how closeout is recorded.
7.6.16(h) Where evidence has multiple classifications, including public authority, health-sensitive, cyber-sensitive, infrastructure-sensitive, finance-sensitive, commercially sensitive, personal information, community-protected, protected knowledge, or controlled technology classifications, GCRI Canada shall apply the most restrictive lawful and public-safe handling unless a competent review records a narrower permissible treatment.
7.6.16(i) Classification shall be reviewed when evidence is corrected, repurposed, routed, published, summarized, translated, modeled, dashboarded, mapped, used in public authority learning, used in finance-facing materials, used in procurement-facing materials, used in provider-facing materials, used in sponsor-facing materials, used in community-facing materials, or archived.
7.6.16(j) The controlling rule shall be that evidence classification is not a label of convenience; it is the operating rule for who may see evidence, how it may be used, how long it may be retained, how it may be communicated, and how it must be corrected.
7.7 Confidence, Uncertainty, and Fitness for Purpose
7.7.1 Confidence as Method-Based Evidence Attribute. 7.7.1(a) Confidence shall be a method-based evidence attribute within the Evidence Rail and shall express the recorded degree to which evidence may be relied upon for a defined purpose, within a defined scope, under defined conditions, subject to recorded limitations, review status, and correction path.
7.7.1(b) Confidence shall not be assigned as institutional optimism, reputational reassurance, sponsor-facing language, provider-facing language, public authority-facing comfort, finance-facing signal, public communications polish, or executive preference. Confidence shall arise from source quality, method quality, data quality, corroboration, independence, calibration, validation, review, completeness, timeliness, reproducibility where applicable, and limitation treatment.
7.7.1(c) Confidence records shall identify, where material, the evidence class, source lineage, provenance, custody, method, data quality, collection conditions, system boundary, calibration status, validation status, reviewer status, conflict status, public-safe status, and correction history.
7.7.1(d) Confidence may be expressed through controlled vocabulary, qualitative levels, structured notes, numerical scores where appropriate, confidence intervals where statistically supported, confidence bands, uncertainty classes, fit-for-purpose findings, or evidence-quality labels approved for use by GCRI Canada.
7.7.1(e) Confidence shall be domain-sensitive. Confidence treatment for AI outputs, cyber logs, sensor records, AI-RAN signals, DePIN records, blockchain anchors, public authority inputs, community knowledge, health-sensitive evidence, geospatial evidence, digital twin outputs, finance-facing inputs, and provider-supplied materials shall reflect the specific risks and limitations of each evidence type.
7.7.1(f) Confidence shall not create certification, recognition, finance-readiness, public authority approval, public warning, emergency command, procurement approval, provider endorsement, sponsor approval, rating, guarantee, protocol effect, operational clearance, or execution consequence.
7.7.1(g) Where confidence is unclear, unsupported, disputed, derived from weak sources, based on incomplete records, or vulnerable to public misunderstanding, GCRI Canada shall classify the evidence conservatively, limit claims, require review, restrict publication, or withhold use until adequate confidence treatment is recorded.
7.7.1(h) The controlling rule shall be that confidence is an evidence-quality attribute created by method and record, not by institutional desire or narrative force.
7.7.2 Uncertainty as Required Disclosure for Material Evidence. 7.7.2(a) Uncertainty shall be a required disclosure for material evidence where uncertainty affects interpretation, use, public-safe publication, public authority learning, finance-facing treatment, procurement-facing risk, provider-facing claims, sponsor-facing materials, community-facing communication, technical baseline support, Docket inputs, Grid inputs, Observatory records, Truth Engine methods, Rails handoffs, or public claims.
7.7.2(b) Uncertainty may include measurement uncertainty, source uncertainty, method uncertainty, model uncertainty, AI uncertainty, geospatial uncertainty, temporal uncertainty, sensor uncertainty, calibration uncertainty, validation uncertainty, classification uncertainty, legal uncertainty, public authority uncertainty, community context uncertainty, finance-facing uncertainty, and public-safe uncertainty.
7.7.2(c) Uncertainty records shall identify what is uncertain, why it is uncertain, how material the uncertainty is, whether the uncertainty can be reduced, what additional evidence would reduce it, what uses remain permitted, what uses are prohibited, and what correction or re-review triggers apply.
7.7.2(d) Uncertainty shall not be concealed by formatting, dashboards, maps, charts, executive summaries, AI-generated summaries, public authority proximity, sponsor framing, provider framing, certainty language, urgency language, or visual design.
7.7.2(e) Public-facing evidence shall disclose uncertainty in a manner proportionate to audience, risk, and public-safe needs. Where full technical uncertainty cannot be safely or clearly disclosed, GCRI Canada shall use public-safe summaries, controlled annexes, limitation statements, or restricted disclosure.
7.7.2(f) Uncertainty disclosure shall not be used to paralyze public-benefit learning where evidence remains fit for a bounded purpose. It shall be used to prevent unsupported authority, false precision, unsafe reliance, inflated claims, and misuse beyond the record.
7.7.2(g) Where uncertainty is material and unresolved, evidence shall not be used for claims of proof, validation, maturity, readiness, certification, recognition, finance-readiness, public authority approval, procurement readiness, provider superiority, safety, security, resilience, or guarantee.
7.7.2(h) The controlling rule shall be that uncertainty is not a weakness to hide; it is part of the evidence that makes the evidence honest.
7.7.3 Confidence Scores as Evidence-Quality Signals, Not Ratings, Certifications, Public Authority Findings, or Finance Determinations. 7.7.3(a) Confidence scores, confidence bands, quality labels, evidence grades, reliability indicators, uncertainty classes, fit-for-purpose findings, readiness-context labels, maturity-context labels, and similar evidence-quality signals shall be used only as evidence-quality signals within the Evidence Rail unless a separate competent authority and record lawfully creates another effect.
7.7.3(b) Confidence scores shall not constitute ratings, certifications, recognitions, finance-readiness determinations, insurance ratings, credit ratings, underwriting conclusions, public finance approvals, procurement rankings, provider rankings, public authority findings, public warnings, emergency commands, compliance approvals, safety approvals, security approvals, or guarantees by default.
7.7.3(c) Any confidence score shall identify the scoring method, evidence category, source records, input variables, weighting where any, reviewer where any, uncertainty treatment, limitations, date, version, public-safe status, permitted use, prohibited use, and correction path.
7.7.3(d) Confidence scores shall not be presented with visual design, color coding, badges, seals, rankings, stars, labels, dashboards, map symbols, or public-facing status indicators that imply certification, recognition, public authority approval, finance-readiness, procurement preference, provider endorsement, official warning, or execution consequence.
7.7.3(e) Where confidence scores are used in finance-facing contexts, the record shall include no-investment-advice, no-offer, no-solicitation, no-underwriting, no-rating, no-public-finance-approval, no-guarantee, no-capital-commitment, no-market-operation, and correction language where material.
7.7.3(f) Where confidence scores are used in public authority-facing contexts, the record shall include non-delegation, non-endorsement, non-official-guidance, non-public-authority-decision, non-warning, non-command, non-regulatory, non-procurement, non-public-finance-approval, and correction language where material.