← Back to catalogue
Published

AI Incident Report

vr.wm-ai-010 · wm-ai-010-ai-incident-report

Represent one governed, versioned AI incident reporting case that makes facts, timeline, affected systems, classification, impact, causal assessments, response, obligations, submissions and corrections machine-readable without absorbing the incident or authority masters.

World Models Information and virtual systems INF.AI.INC

Bundle → Layer → Finding → Questions Filled

6 bundles · 12 layers · 24 findings · 72 questions

Report identity, scope, jurisdiction and parties Groups governed report context for report identity, scope, jurisdiction and parties.

Report root, version, status and authority

Groups source-qualified report context for report root, version, status and authority.

Report case identity, type, version, issuer and current head

Records report case identity, type, version, issuer and current head as source-qualified report context while incident, system, investigation, authority, remediation, evidence and records masters remain external.

  1. What stable identity, report version, jurisdiction-qualified value and explicit unknown establish report case identity, type, version, issuer and current head? identity
  2. Who observed, asserted, reviewed, approved, received or may rely on report case identity, type, version, issuer and current head, under which authority and limits? event
  3. Which occurrence, awareness, effective, submitted, recorded, ingested and knowledge times, evidence and uncertainty qualify report case identity, type, version, issuer and current head? process

Jurisdiction, legal basis, recipient, reportability and authority

Records jurisdiction, legal basis, recipient, reportability and authority as source-qualified report context while incident, system, investigation, authority, remediation, evidence and records masters remain external.

  1. What stable identity, report version, jurisdiction-qualified value and explicit unknown establish jurisdiction, legal basis, recipient, reportability and authority? authority
  2. Who observed, asserted, reviewed, approved, received or may rely on jurisdiction, legal basis, recipient, reportability and authority, under which authority and limits? temporal
  3. Which occurrence, awareness, effective, submitted, recorded, ingested and knowledge times, evidence and uncertainty qualify jurisdiction, legal basis, recipient, reportability and authority? validation

Incident, system and party bindings

Groups source-qualified report context for incident, system and party bindings.

Incident, related report, affected system, deployment, model and version

Records incident, related report, affected system, deployment, model and version as source-qualified report context while incident, system, investigation, authority, remediation, evidence and records masters remain external.

  1. What stable identity, report version, jurisdiction-qualified value and explicit unknown establish incident, related report, affected system, deployment, model and version? relationship
  2. Who observed, asserted, reviewed, approved, received or may rely on incident, related report, affected system, deployment, model and version, under which authority and limits? composition
  3. Which occurrence, awareness, effective, submitted, recorded, ingested and knowledge times, evidence and uncertainty qualify incident, related report, affected system, deployment, model and version? privacy

Reporter, provider, deployer, operator, contact and responsibility

Records reporter, provider, deployer, operator, contact and responsibility as source-qualified report context while incident, system, investigation, authority, remediation, evidence and records masters remain external.

  1. What stable identity, report version, jurisdiction-qualified value and explicit unknown establish reporter, provider, deployer, operator, contact and responsibility? ownership
  2. Who observed, asserted, reviewed, approved, received or may rely on reporter, provider, deployer, operator, contact and responsibility, under which authority and limits? evidence
  3. Which occurrence, awareness, effective, submitted, recorded, ingested and knowledge times, evidence and uncertainty qualify reporter, provider, deployer, operator, contact and responsibility? lifecycle
Occurrence, observation, context and timeline Groups governed report context for occurrence, observation, context and timeline.

Event clock, location and evidence source

Groups source-qualified report context for event clock, location and evidence source.

Occurrence, detection, discovery, awareness, notification and report time

Records occurrence, detection, discovery, awareness, notification and report time as source-qualified report context while incident, system, investigation, authority, remediation, evidence and records masters remain external.

  1. What stable identity, report version, jurisdiction-qualified value and explicit unknown establish occurrence, detection, discovery, awareness, notification and report time? temporal
  2. Who observed, asserted, reviewed, approved, received or may rely on occurrence, detection, discovery, awareness, notification and report time, under which authority and limits? ownership
  3. Which occurrence, awareness, effective, submitted, recorded, ingested and knowledge times, evidence and uncertainty qualify occurrence, detection, discovery, awareness, notification and report time? quality

Location, channel, source, observer, confidence, gap and unavailability

Records location, channel, source, observer, confidence, gap and unavailability as source-qualified report context while incident, system, investigation, authority, remediation, evidence and records masters remain external.

  1. What stable identity, report version, jurisdiction-qualified value and explicit unknown establish location, channel, source, observer, confidence, gap and unavailability? evidence
  2. Who observed, asserted, reviewed, approved, received or may rely on location, channel, source, observer, confidence, gap and unavailability, under which authority and limits? measurement
  3. Which occurrence, awareness, effective, submitted, recorded, ingested and knowledge times, evidence and uncertainty qualify location, channel, source, observer, confidence, gap and unavailability? security

Behavior, sequence and operating context

Groups source-qualified report context for behavior, sequence and operating context.

Behavior, failure, malfunction, output, action, outcome and sequence

Records behavior, failure, malfunction, output, action, outcome and sequence as source-qualified report context while incident, system, investigation, authority, remediation, evidence and records masters remain external.

  1. What stable identity, report version, jurisdiction-qualified value and explicit unknown establish behavior, failure, malfunction, output, action, outcome and sequence? event
  2. Who observed, asserted, reviewed, approved, received or may rely on behavior, failure, malfunction, output, action, outcome and sequence, under which authority and limits? exception
  3. Which occurrence, awareness, effective, submitted, recorded, ingested and knowledge times, evidence and uncertainty qualify behavior, failure, malfunction, output, action, outcome and sequence? retention

Lifecycle stage, environment, data, input, output, human and tool context

Records lifecycle stage, environment, data, input, output, human and tool context as source-qualified report context while incident, system, investigation, authority, remediation, evidence and records masters remain external.

  1. What stable identity, report version, jurisdiction-qualified value and explicit unknown establish lifecycle stage, environment, data, input, output, human and tool context? composition
  2. Who observed, asserted, reviewed, approved, received or may rely on lifecycle stage, environment, data, input, output, human and tool context, under which authority and limits? provenance
  3. Which occurrence, awareness, effective, submitted, recorded, ingested and knowledge times, evidence and uncertainty qualify lifecycle stage, environment, data, input, output, human and tool context? interoperability
Classification, severity, impact and harm Groups governed report context for classification, severity, impact and harm.

Incident class, severity and causal threshold

Groups source-qualified report context for incident class, severity and causal threshold.

Incident, hazard, near miss, serious criterion, category and severity

Records incident, hazard, near miss, serious criterion, category and severity as source-qualified report context while incident, system, investigation, authority, remediation, evidence and records masters remain external.

  1. What stable identity, report version, jurisdiction-qualified value and explicit unknown establish incident, hazard, near miss, serious criterion, category and severity? classification
  2. Who observed, asserted, reviewed, approved, received or may rely on incident, hazard, near miss, serious criterion, category and severity, under which authority and limits? process
  3. Which occurrence, awareness, effective, submitted, recorded, ingested and knowledge times, evidence and uncertainty qualify incident, hazard, near miss, serious criterion, category and severity? decision

Causal link, likelihood, attribution, scope, confidence and uncertainty

Records causal link, likelihood, attribution, scope, confidence and uncertainty as source-qualified report context while incident, system, investigation, authority, remediation, evidence and records masters remain external.

  1. What stable identity, report version, jurisdiction-qualified value and explicit unknown establish causal link, likelihood, attribution, scope, confidence and uncertainty? decision
  2. Who observed, asserted, reviewed, approved, received or may rely on causal link, likelihood, attribution, scope, confidence and uncertainty, under which authority and limits? validation
  3. Which occurrence, awareness, effective, submitted, recorded, ingested and knowledge times, evidence and uncertainty qualify causal link, likelihood, attribution, scope, confidence and uncertainty? state

Affected subjects, rights, assets and consequences

Groups source-qualified report context for affected subjects, rights, assets and consequences.

Person, group, right, health, safety, property, infrastructure and environment

Records person, group, right, health, safety, property, infrastructure and environment as source-qualified report context while incident, system, investigation, authority, remediation, evidence and records masters remain external.

  1. What stable identity, report version, jurisdiction-qualified value and explicit unknown establish person, group, right, health, safety, property, infrastructure and environment? measurement
  2. Who observed, asserted, reviewed, approved, received or may rely on person, group, right, health, safety, property, infrastructure and environment, under which authority and limits? privacy
  3. Which occurrence, awareness, effective, submitted, recorded, ingested and knowledge times, evidence and uncertainty qualify person, group, right, health, safety, property, infrastructure and environment? identity

Geography, system, service, scale, duration, persistence and reversibility

Records geography, system, service, scale, duration, persistence and reversibility as source-qualified report context while incident, system, investigation, authority, remediation, evidence and records masters remain external.

  1. What stable identity, report version, jurisdiction-qualified value and explicit unknown establish geography, system, service, scale, duration, persistence and reversibility? measurement
  2. Who observed, asserted, reviewed, approved, received or may rely on geography, system, service, scale, duration, persistence and reversibility, under which authority and limits? lifecycle
  3. Which occurrence, awareness, effective, submitted, recorded, ingested and knowledge times, evidence and uncertainty qualify geography, system, service, scale, duration, persistence and reversibility? classification
Cause, security, privacy, investigation and evidence Groups governed report context for cause, security, privacy, investigation and evidence.

Cause, contributor, investigation and assessment

Groups source-qualified report context for cause, contributor, investigation and assessment.

Model, data, software, hardware, supplier, human, process and organization factor

Records model, data, software, hardware, supplier, human, process and organization factor as source-qualified report context while incident, system, investigation, authority, remediation, evidence and records masters remain external.

  1. What stable identity, report version, jurisdiction-qualified value and explicit unknown establish model, data, software, hardware, supplier, human, process and organization factor? provenance
  2. Who observed, asserted, reviewed, approved, received or may rely on model, data, software, hardware, supplier, human, process and organization factor, under which authority and limits? quality
  3. Which occurrence, awareness, effective, submitted, recorded, ingested and knowledge times, evidence and uncertainty qualify model, data, software, hardware, supplier, human, process and organization factor? relationship

Root or contributing factor, hypothesis, investigation method, finding and review

Records root or contributing factor, hypothesis, investigation method, finding and review as source-qualified report context while incident, system, investigation, authority, remediation, evidence and records masters remain external.

  1. What stable identity, report version, jurisdiction-qualified value and explicit unknown establish root or contributing factor, hypothesis, investigation method, finding and review? validation
  2. Who observed, asserted, reviewed, approved, received or may rely on root or contributing factor, hypothesis, investigation method, finding and review, under which authority and limits? security
  3. Which occurrence, awareness, effective, submitted, recorded, ingested and knowledge times, evidence and uncertainty qualify root or contributing factor, hypothesis, investigation method, finding and review? authority

Security, adversarial and personal-data breach

Groups source-qualified report context for security, adversarial and personal-data breach.

Threat actor, vulnerability, attack technique, poisoning, evasion, privacy and misuse

Records threat actor, vulnerability, attack technique, poisoning, evasion, privacy and misuse as source-qualified report context while incident, system, investigation, authority, remediation, evidence and records masters remain external.

  1. What stable identity, report version, jurisdiction-qualified value and explicit unknown establish threat actor, vulnerability, attack technique, poisoning, evasion, privacy and misuse? security
  2. Who observed, asserted, reviewed, approved, received or may rely on threat actor, vulnerability, attack technique, poisoning, evasion, privacy and misuse, under which authority and limits? retention
  3. Which occurrence, awareness, effective, submitted, recorded, ingested and knowledge times, evidence and uncertainty qualify threat actor, vulnerability, attack technique, poisoning, evasion, privacy and misuse? requirement

Personal-data breach, confidentiality, integrity, availability, risk and notification

Records personal-data breach, confidentiality, integrity, availability, risk and notification as source-qualified report context while incident, system, investigation, authority, remediation, evidence and records masters remain external.

  1. What stable identity, report version, jurisdiction-qualified value and explicit unknown establish personal-data breach, confidentiality, integrity, availability, risk and notification? privacy
  2. Who observed, asserted, reviewed, approved, received or may rely on personal-data breach, confidentiality, integrity, availability, risk and notification, under which authority and limits? interoperability
  3. Which occurrence, awareness, effective, submitted, recorded, ingested and knowledge times, evidence and uncertainty qualify personal-data breach, confidentiality, integrity, availability, risk and notification? constraint
Response, remediation, notification and authority exchange Groups governed report context for response, remediation, notification and authority exchange.

Containment, recovery, corrective and preventive action

Groups source-qualified report context for containment, recovery, corrective and preventive action.

Immediate containment, safeguard, continuity, recovery and residual risk

Records immediate containment, safeguard, continuity, recovery and residual risk as source-qualified report context while incident, system, investigation, authority, remediation, evidence and records masters remain external.

  1. What stable identity, report version, jurisdiction-qualified value and explicit unknown establish immediate containment, safeguard, continuity, recovery and residual risk? process
  2. Who observed, asserted, reviewed, approved, received or may rely on immediate containment, safeguard, continuity, recovery and residual risk, under which authority and limits? decision
  3. Which occurrence, awareness, effective, submitted, recorded, ingested and knowledge times, evidence and uncertainty qualify immediate containment, safeguard, continuity, recovery and residual risk? event

Corrective or preventive action, owner, due date, verification and effectiveness

Records corrective or preventive action, owner, due date, verification and effectiveness as source-qualified report context while incident, system, investigation, authority, remediation, evidence and records masters remain external.

  1. What stable identity, report version, jurisdiction-qualified value and explicit unknown establish corrective or preventive action, owner, due date, verification and effectiveness? lifecycle
  2. Who observed, asserted, reviewed, approved, received or may rely on corrective or preventive action, owner, due date, verification and effectiveness, under which authority and limits? state
  3. Which occurrence, awareness, effective, submitted, recorded, ingested and knowledge times, evidence and uncertainty qualify corrective or preventive action, owner, due date, verification and effectiveness? temporal

Obligation, deadline, submission and authority response

Groups source-qualified report context for obligation, deadline, submission and authority response.

Recipient, regime, trigger, awareness clock, deadline, exception and extension

Records recipient, regime, trigger, awareness clock, deadline, exception and extension as source-qualified report context while incident, system, investigation, authority, remediation, evidence and records masters remain external.

  1. What stable identity, report version, jurisdiction-qualified value and explicit unknown establish recipient, regime, trigger, awareness clock, deadline, exception and extension? requirement
  2. Who observed, asserted, reviewed, approved, received or may rely on recipient, regime, trigger, awareness clock, deadline, exception and extension, under which authority and limits? identity
  3. Which occurrence, awareness, effective, submitted, recorded, ingested and knowledge times, evidence and uncertainty qualify recipient, regime, trigger, awareness clock, deadline, exception and extension? composition

Initial, intermediate or final submission, receipt, question, response and status

Records initial, intermediate or final submission, receipt, question, response and status as source-qualified report context while incident, system, investigation, authority, remediation, evidence and records masters remain external.

  1. What stable identity, report version, jurisdiction-qualified value and explicit unknown establish initial, intermediate or final submission, receipt, question, response and status? relationship
  2. Who observed, asserted, reviewed, approved, received or may rely on initial, intermediate or final submission, receipt, question, response and status, under which authority and limits? classification
  3. Which occurrence, awareness, effective, submitted, recorded, ingested and knowledge times, evidence and uncertainty qualify initial, intermediate or final submission, receipt, question, response and status? evidence
Governance, disclosure, lifecycle, records and projections Groups governed report context for governance, disclosure, lifecycle, records and projections.

Approval, confidentiality, communication and public summary

Groups source-qualified report context for approval, confidentiality, communication and public summary.

Author, approver, legal review, attestation, conflict and accountability

Records author, approver, legal review, attestation, conflict and accountability as source-qualified report context while incident, system, investigation, authority, remediation, evidence and records masters remain external.

  1. What stable identity, report version, jurisdiction-qualified value and explicit unknown establish author, approver, legal review, attestation, conflict and accountability? authority
  2. Who observed, asserted, reviewed, approved, received or may rely on author, approver, legal review, attestation, conflict and accountability, under which authority and limits? relationship
  3. Which occurrence, awareness, effective, submitted, recorded, ingested and knowledge times, evidence and uncertainty qualify author, approver, legal review, attestation, conflict and accountability? ownership

Confidentiality, privilege, trade secret, minimization, public summary and notice

Records confidentiality, privilege, trade secret, minimization, public summary and notice as source-qualified report context while incident, system, investigation, authority, remediation, evidence and records masters remain external.

  1. What stable identity, report version, jurisdiction-qualified value and explicit unknown establish confidentiality, privilege, trade secret, minimization, public summary and notice? privacy
  2. Who observed, asserted, reviewed, approved, received or may rely on confidentiality, privilege, trade secret, minimization, public summary and notice, under which authority and limits? authority
  3. Which occurrence, awareness, effective, submitted, recorded, ingested and knowledge times, evidence and uncertainty qualify confidentiality, privilege, trade secret, minimization, public summary and notice? measurement

Evidence, correction, retention, audit and interoperability

Groups source-qualified report context for evidence, correction, retention, audit and interoperability.

Attachment, evidence, digest, chain of custody, source and provenance

Records attachment, evidence, digest, chain of custody, source and provenance as source-qualified report context while incident, system, investigation, authority, remediation, evidence and records masters remain external.

  1. What stable identity, report version, jurisdiction-qualified value and explicit unknown establish attachment, evidence, digest, chain of custody, source and provenance? evidence
  2. Who observed, asserted, reviewed, approved, received or may rely on attachment, evidence, digest, chain of custody, source and provenance, under which authority and limits? requirement
  3. Which occurrence, awareness, effective, submitted, recorded, ingested and knowledge times, evidence and uncertainty qualify attachment, evidence, digest, chain of custody, source and provenance? exception

Draft, submitted, accepted, amended, superseded, closed, reopened, retained and projected

Records draft, submitted, accepted, amended, superseded, closed, reopened, retained and projected as source-qualified report context while incident, system, investigation, authority, remediation, evidence and records masters remain external.

  1. What stable identity, report version, jurisdiction-qualified value and explicit unknown establish draft, submitted, accepted, amended, superseded, closed, reopened, retained and projected? lifecycle
  2. Who observed, asserted, reviewed, approved, received or may rely on draft, submitted, accepted, amended, superseded, closed, reopened, retained and projected, under which authority and limits? constraint
  3. Which occurrence, awareness, effective, submitted, recorded, ingested and knowledge times, evidence and uncertainty qualify draft, submitted, accepted, amended, superseded, closed, reopened, retained and projected? provenance

Classifiers Filled

Family
World Models
Category
Information and virtual systems
Entry kind
aggregate
Navigation path
NAV.INF.AI.INC
Domain
INF.AI.INC
Industry
Cross-industry
Tags
aiincidentreportinf.ai.inc

What it is Filled

Owns report-case identity, version and status; issuer, jurisdiction, legal basis, recipient and reportability assertions; incident, affected-system and party bindings; facts, timeline, context, classification, severity, impact, harm, causal, security and privacy assessments; response and remediation summaries; notification duties, deadlines, submissions, receipts and authority exchanges; approval, confidentiality, communication, evidence, correction, retention, audit and projections. Incident, AI system, model, deployment, investigation, regulatory case, remediation work, harm, vulnerability, evidence, audit and records masters remain external.

In scope

  • Report identity, version, jurisdiction, legal basis, recipient, affected systems, responsible parties, facts, timeline, context, classification, severity, impact and causal assessments
  • Security and privacy links, response and remediation summaries, obligations, deadlines, submissions, receipts, authority exchanges, disclosure, lifecycle, evidence, retention, audit and projections

Out of scope

  • Owning the incident, affected AI system, model, deployment, person, organization, harm, investigation, regulatory case, remediation action, vulnerability, evidence, audit or records masters
  • Treating report receipt, mitigation deployment, system recovery or closure as proof of causality, acceptance, effectiveness, compliance or absence of harm
  • Autonomous legal classification, regulatory submission, public disclosure, liability admission, attribution, access expansion, evidence destruction or closure

Why it exists Filled

Represent one governed, versioned AI incident reporting case that makes facts, timeline, affected systems, classification, impact, causal assessments, response, obligations, submissions and corrections machine-readable without absorbing the incident or authority masters.

Distinguishing features Filled

  • A reporting case with versions and submissions about an incident, not the incident itself.
  • Keeps observed fact, reporter statement, hypothesis, causal assessment and authority finding apart.
  • Does not equate an AI incident with a security incident, data breach or defect, though it may link them.
  • Carries reportability, recipient and deadline assertions tied to a legal basis.

What robots and AI may and may not do Filled

Must not

  • Submit a report to an authority.
  • Make a binding legal classification or reportability decision.
  • Admit liability or assign blame to a person.
  • Disclose restricted details or personal data of affected people.
  • Alter facts in a submitted version instead of issuing a correction.

Only with a human decision

  • Deciding whether an incident is reportable and to whom.
  • Approving and submitting a report.
  • Public disclosure of incident details.

May

  • Collect source-qualified facts and build a timeline.
  • Calculate reporting deadlines from the declared legal basis.
  • Draft initial, intermediate and final report versions for review.
  • Link related security, privacy and investigation records.

Moral aspects Filled

  • Affected people deserve timely notice and remedy, so delays in reporting cause real harm.
  • Reports must protect the identity of affected people and of whistleblowers.
  • Blame on individual operators can hide systemic causes in design or deployment.

Who is affected

  • People harmed or affected by the AI system
  • Reporters and whistleblowers
  • Deployers, providers and authorities

Owners Filled

Steward

Dimension owner, incident-reporting mandate and accountable AI system owner

Roles

AI system owner and accountable deployer
Own reporting readiness, accountable assessment, response coordination and use of incident reports.
Incident reporter and case coordinator
Maintain report identity, source-qualified facts, timeline, versions, submissions and authority exchanges.
Technical investigator and AI specialist
Supply bounded system, model, data, failure, causal and corrective evidence with uncertainty.
Safety, security and privacy responder
Assess harm, threat, vulnerability, privacy-breach, containment, notification and residual risk profiles.
Legal, compliance and regulatory liaison
Determine applicable regimes, deadlines, recipients, privilege, confidentiality and approved submissions.
Independent reviewer and approval authority
Review evidence, conflicts, classifications, causal claims, completeness and release decisions within mandate.
Records and disclosure steward
Own access, minimization, public summaries, legal hold, retention, disposition and projection policy.

Links to other meta-models Filled

references

  • WM-AI-001 AI System - Represent the unfrozen parent and affected-system boundary without inheritance, composition, mutation, release or cascade authority.
  • WM-ACT-019 Incident - Resolve the underlying general incident while preserving a separate reporting-case identity and lifecycle.
  • Investigation, regulatory case, remediation, harm, vulnerability, evidence, audit and records models - Resolve authoritative findings, actions, decisions and evidence without absorbing their ownership.

aligned

  • EU AI Act, OECD AI incident framework, NIST AI RMF and SP 800-61r3, FIRST CSIRT, ISO/IEC 42001, PROV-O and RFC 3339 - Project version-pinned legal, reporting, risk, response, management, provenance and timestamp views with declared loss.

neighbor

  • WM-AI-001 AI System - The unified parent signal and candidate reference identify affected-system context but grant no inheritance or ownership. The report binds source-qualified system, model and deployment versions without mutating their masters.
  • WM-ACT-019 Incident - The general incident master owns occurrence, operational state and response history. The report owns a versioned communication and submission case about that incident.
  • Investigation, remediation and regulatory case - Investigations own methods and findings, remediation work owns actions and verification, and authorities own their cases and decisions. The report stores non-owning summaries, submissions, receipts and exchanges.
  • Security incident and personal-data breach - Cybersecurity and privacy regimes may overlap an AI incident but use independent applicability, risk, notification and authority semantics.
  • EU AI Act, OECD, NIST, FIRST, ISO, PROV and RFC profiles - These legal, reporting, risk, incident-response, management, provenance and timestamp sources have distinct scopes. No mapping is assumed binding or lossless outside its profile.

parent

  • WM-AI-001

What else AI and robots need to interact with it Filled

Identity and identifiers required Filled

  • Authoritative master-system identifier for each report case, version, submission, assertion, decision or projection, qualified by issuer, namespace and record kind.
  • Governed globally resolvable report-case IRI.
  • Dimension UUID or ULID when neither preceding identifier exists.

Direct properties not applicable Not applicable

Not applicable

Institutional or informational subject: no invented physical properties.

Recognition optional Filled

  • An incident report names an affected AI system, an event time, a harm or near miss, a reporter and a submission state.
  • Often confused with a security incident ticket, a data breach notification, a customer complaint or a bug report.

Capabilities and actions required Filled

  • Open an AI incident report case: Governed operation to open an ai incident report case without autonomous legal classification, regulatory submission, public disclosure, liability admission, attribution, access expansion, evidence destruction or closure.
  • Bind affected systems and responsible parties: Governed operation to bind affected systems and responsible parties without autonomous legal classification, regulatory submission, public disclosure, liability admission, attribution, access expansion, evidence destruction or closure.
  • Record source-qualified facts and timeline: Governed operation to record source-qualified facts and timeline without autonomous legal classification, regulatory submission, public disclosure, liability admission, attribution, access expansion, evidence destruction or closure.
  • Assess classification, impact and causality: Governed operation to assess classification, impact and causality without autonomous legal classification, regulatory submission, public disclosure, liability admission, attribution, access expansion, evidence destruction or closure.
  • Link security, privacy and investigation records: Governed operation to link security, privacy and investigation records without autonomous legal classification, regulatory submission, public disclosure, liability admission, attribution, access expansion, evidence destruction or closure.
  • Record response and remediation summaries: Governed operation to record response and remediation summaries without autonomous legal classification, regulatory submission, public disclosure, liability admission, attribution, access expansion, evidence destruction or closure.
  • Calculate a reporting obligation and deadline: Governed operation to calculate a reporting obligation and deadline without autonomous legal classification, regulatory submission, public disclosure, liability admission, attribution, access expansion, evidence destruction or closure.
  • Submit an initial, intermediate, final or corrected report: Governed operation to submit an initial, intermediate, final or corrected report without autonomous legal classification, regulatory submission, public disclosure, liability admission, attribution, access expansion, evidence destruction or closure.
  • Amend, supersede, close or reopen a report: Governed operation to amend, supersede, close or reopen a report without autonomous legal classification, regulatory submission, public disclosure, liability admission, attribution, access expansion, evidence destruction or closure.
  • Disclose, project, retain and audit: Governed operation to disclose, project, retain and audit without autonomous legal classification, regulatory submission, public disclosure, liability admission, attribution, access expansion, evidence destruction or closure.

Hazards and failure modes required Filled

  • Missed legal reporting deadlines.
  • Leaks of personal or sensitive data through reports.
  • Premature causal claims that misdirect remediation.

Standards and interfaces required Filled

  • EU AI Act (Regulation (EU) 2024/1689) Article 73 serious incident reporting.
  • NIST AI Risk Management Framework 1.0.
  • ISO/IEC 42001 AI management systems.
  • OASIS STIX 2.1 for linked security incidents.
  • RFC 3339 timestamps.
  • W3C PROV-O.

Context of use required Filled

  • Incident definitions, reportability, seriousness, causal threshold, deadlines, recipients, privilege, disclosure, privacy and retention depend on jurisdiction and sector.
  • The EU AI Act, GDPR and EDPB guidance are European Union profiles; NIST publications are voluntary United States guidance unless adopted.
  • OECD, FIRST, ISO, PROV and RFC sources are versioned profiles and are not assumed binding, lossless or universally applicable.

Sources Filled

  1. Regulation (EU) 2024/1689 Artificial Intelligence Act, consolidated text - European Union
  2. Draft guidance and reporting template on serious AI incidents - European Commission
  3. Defining AI incidents and related terms - Organisation for Economic Co-operation and Development
  4. Towards a common reporting framework for AI incidents - Organisation for Economic Co-operation and Development
  5. AI Incidents and Hazards Monitor - Organisation for Economic Co-operation and Development
  6. Artificial Intelligence Risk Management Framework (AI RMF 1.0) - National Institute of Standards and Technology
  7. Artificial Intelligence Risk Management Framework: Generative Artificial Intelligence Profile - National Institute of Standards and Technology
  8. Incident Response Recommendations and Considerations for Cybersecurity Risk Management - National Institute of Standards and Technology
  9. Adversarial Machine Learning: A Taxonomy and Terminology of Attacks and Mitigations - National Institute of Standards and Technology
  10. Regulation (EU) 2016/679 General Data Protection Regulation - European Union
  11. Guidelines 9/2022 on personal data breach notification under GDPR - European Data Protection Board
  12. CSIRT Services Framework - Forum of Incident Response and Security Teams
  13. ISO/IEC 42001 AI management systems - International Organization for Standardization
  14. PROV-O: The PROV Ontology - World Wide Web Consortium
  15. Date and Time on the Internet: Timestamps - Internet Engineering Task Force

Open questions

  • Approve or reject the WM-AI-001 and WM-ACT-019 relationships and register investigation, remediation, authority-case, harm, vulnerability, evidence and records edges.
  • Create jurisdiction, sector, AI-system-class, seriousness, causal-threshold, deadline, recipient and reporting-template profiles.
  • Refresh the EU Article 73 mapping when final guidance and templates replace the inspected draft package.
  • Validate organization-specific privilege, trade-secret, personal-data, security, disclosure, affected-person communication and retention policies.
  • Test version-pinned EU, OECD, NIST, FIRST, ISO, PROV and RFC mappings and obtain supplemental independent external review before canonical promotion.
  • Claude and Grok each timed out on one bounded attempt; no independent external result was admitted.
  • The unified parent WM-AI-001 and candidate references to WM-AI-001 and WM-ACT-019 are not approved frozen composition edges.
  • The European Commission Article 73 guidance and high-risk-system template inspected here remain draft.
  • Sector, jurisdiction, system-class and regulator-specific notification forms and deadlines require separate profiles.

Machine files

Provenance

world-models research · reviewable-draft

Built from: models/wm-ai-010-ai-incident-report/spec.yaml, ver-cy/world-models/card-supplements/wm-ai-010-ai-incident-report.json