Narrative Goldmine

etsi domain taxonomy

49 min read

Properties

Type
  • Class
Resource
  • urn:ngm:class:etsi-domain-taxonomy
Domain
  • artificial-intelligence
Status
  • stable
Maturity
  • emerging
Quality
  • 0.72
Is a
Part of
Has part
Requires
Enables
Generated
  • by: process:vault-migrate/1.0 · at: 2026-09-22T12:41:39.024503178Z
Bridges to
Contrasts with
Depends on
Implemented in layer
Implements
Label
  • ETSI Domain Taxonomy
Links
    Page resource
    • urn:visionflow:page:8a4313c0d8a7a59ac4839449416255bc5dd6bdf149014fcaee2244361c1ec17e
    Related to
    Standardized by
    Supports
    Uses

    The ETSI Domain Taxonomy is a hierarchical classification framework produced by the European Telecommunications Standards Institute for systematically organising AI-related technical specifications, standards, and work items across application domains such as transportation, healthcare, telecommunications, and public safety. It provides a controlled vocabulary and structured ontology enabling consistent referencing of AI standards across ETSI technical committees and supports cross-domain applicability analysis. The taxonomy underpins ETSI’s AI standardisation roadmap and aligns with CEN-CENELEC and ISO/IEC JTC 1/SC 42 standardisation activities. As a living classification instrument, it maps broad application sectors down to specific sub-domains, allowing AI systems and standards to receive multiple domain affiliations and enabling gap analysis across the European standardisation ecosystem.

    Semantic Classification

    Content

    Compositional Relationships (Components)

    SubClassOf(ai:ETSIDomainTaxonomy
  ObjectSomeValuesFrom(ai:hasPart ai:ETSIDomainAI))
SubClassOf(ai:ETSIDomainTaxonomy
  ObjectSomeValuesFrom(ai:hasPart ai:ETSIDomainData))
SubClassOf(ai:ETSIDomainTaxonomy
  ObjectSomeValuesFrom(ai:hasPart ai:ETSIDomainSecurityAndPrivacy))
SubClassOf(ai:ETSIDomainTaxonomy
  ObjectSomeValuesFrom(ai:hasPart ai:ETSIDomainIdentityAndTrust))
SubClassOf(ai:ETSIDomainTaxonomy
  ObjectSomeValuesFrom(ai:hasPart ai:AIApplicationDomain))
SubClassOf(ai:ETSIDomainTaxonomy
  ObjectSomeValuesFrom(ai:hasPart ai:ControlledVocabulary))
SubClassOf(ai:ETSIDomainTaxonomy
  ObjectSomeValuesFrom(ai:hasPart ai:DomainClassificationHierarchy))

    Dependency Relationships

    SubClassOf(ai:ETSIDomainTaxonomy
  ObjectSomeValuesFrom(ai:requires ai:TechnicalCommittee))
SubClassOf(ai:ETSIDomainTaxonomy
  ObjectSomeValuesFrom(ai:requires ai:TaxonomyGovernance))
SubClassOf(ai:ETSIDomainTaxonomy
  ObjectSomeValuesFrom(ai:dependsOn ai:ETSI))
SubClassOf(ai:ETSIDomainTaxonomy
  ObjectSomeValuesFrom(ai:dependsOn ai:ISOIECJtc1Sc42))
SubClassOf(ai:ETSIDomainTaxonomy
  ObjectSomeValuesFrom(ai:dependsOn ai:CenCenelec))
SubClassOf(ai:ETSIDomainTaxonomy
  ObjectSomeValuesFrom(ai:uses ai:Ontology))
SubClassOf(ai:ETSIDomainTaxonomy
  ObjectSomeValuesFrom(ai:uses ai:SemanticWeb))
SubClassOf(ai:ETSIDomainTaxonomy
  ObjectSomeValuesFrom(ai:uses ai:KnowledgeRepresentation))

    Capability Relationships

    SubClassOf(ai:ETSIDomainTaxonomy
  ObjectSomeValuesFrom(ai:enables ai:StandardsGapAnalysis))
SubClassOf(ai:ETSIDomainTaxonomy
  ObjectSomeValuesFrom(ai:enables ai:CrossDomainApplicability))
SubClassOf(ai:ETSIDomainTaxonomy
  ObjectSomeValuesFrom(ai:enables ai:AIConformityAssessment))
SubClassOf(ai:ETSIDomainTaxonomy
  ObjectSomeValuesFrom(ai:enables ai:AIServiceClassification))
SubClassOf(ai:ETSIDomainTaxonomy
  ObjectSomeValuesFrom(ai:supports ai:EUAIAct))
SubClassOf(ai:ETSIDomainTaxonomy
  ObjectSomeValuesFrom(ai:supports ai:HarmonisedStandards))
SubClassOf(ai:ETSIDomainTaxonomy
  ObjectSomeValuesFrom(ai:supports ai:Interoperability))

    Implementation Relationships

    SubClassOf(ai:ETSIDomainTaxonomy
  ObjectSomeValuesFrom(ai:implements ai:AIStandardisation))
SubClassOf(ai:ETSIDomainTaxonomy
  ObjectSomeValuesFrom(ai:implements ai:ClassificationFramework))
SubClassOf(ai:ETSIDomainTaxonomy
  ObjectSomeValuesFrom(ai:standardizedBy ai:ETSI))
SubClassOf(ai:ETSIDomainTaxonomy
  ObjectSomeValuesFrom(ai:standardizedBy ai:ISOIECJtc1Sc42))

    Reduction Relationships

    SubClassOf(ai:ETSIDomainTaxonomy
  ObjectSomeValuesFrom(ai:reducesTo ai:AIGovernanceAndEthics))
SubClassOf(ai:ETSIDomainTaxonomy
  ObjectSomeValuesFrom(ai:reducesTo ai:KnowledgeOrganisationSystem))
SubClassOf(ai:ETSIDomainTaxonomy
  ObjectSomeValuesFrom(ai:reducesTo ai:ControlledVocabulary))
SubClassOf(ai:ETSIDomainTaxonomy
  ObjectSomeValuesFrom(ai:reducesTo ai:StandardisationFramework))
SubClassOf(ai:ETSIDomainTaxonomy
  ObjectSomeValuesFrom(ai:relatedTo ai:AIGovernanceAndEthics))
SubClassOf(ai:ETSIDomainTaxonomy
  ObjectSomeValuesFrom(ai:relatedTo ai:EuropeanAIStandardsEcosystem))
SubClassOf(ai:ETSIDomainTaxonomy
  ObjectSomeValuesFrom(ai:contrastsWith ai:ISOIEC42001))
SubClassOf(ai:ETSIDomainTaxonomy
  ObjectSomeValuesFrom(ai:contrastsWith ai:NISTAIRiskManagementFramework))
SubClassOf(ai:ETSIDomainTaxonomy
  ObjectSomeValuesFrom(ai:bridges ai:MachineLearningOperations))
SubClassOf(ai:ETSIDomainTaxonomy
  ObjectSomeValuesFrom(ai:bridges ai:DataGovernance))
SubClassOf(ai:ETSIDomainTaxonomy
  ObjectSomeValuesFrom(ai:relatedTo ai:ETSIISGENIEG))
SubClassOf(ai:ETSIDomainTaxonomy
  ObjectSomeValuesFrom(ai:relatedTo ai:ETSITCSAIGroup))
SubClassOf(ai:ETSIDomainTaxonomy
  ObjectSomeValuesFrom(ai:relatedTo ai:ETSIISGMECGroup))
SubClassOf(ai:ETSIDomainTaxonomy
  ObjectSomeValuesFrom(ai:implements ai:EuropeanAIActHarmonisedStandards))

    About

    The ETSI Domain Taxonomy emerged from a structural recognition within ETSI that its AI-related standardisation work had proliferated across too many distinct Technical Committees and Industry Specification Groups to be navigated coherently without a shared classification framework. By the early 2020s, ETSI counted 14 or more technical bodies producing AI-relevant specifications, from the ISG ENI (Experiential Networked Intelligence) addressing cognitive network automation, to the ISG SAI (later TC SAI) addressing AI security, ISG MEC addressing edge computing with AI enablers, ISG ZSM addressing zero-touch network management, and TC Cyber addressing cybersecurity specifications with AI components. Without a unifying taxonomy, the risk of conflicting definitions, overlapping scope claims, and gaps in coverage was significant. The ETSI Domain Taxonomy resolves this by providing a canonical set of domain classification nodes — including but not limited to ETSI Domain AI, ETSI Domain Data, ETSI Domain Security and Privacy, and cross-domain intersections such as ETSI Domain AI + Governance and ETSI Domain AI + Data Mgmt — against which all ETSI work items in the AI space can be tagged, enabling systematic gap analysis and consistent cross-referencing.

    The taxonomy’s design philosophy reflects two sometimes-competing imperatives. First, it must be granular enough to capture genuine distinctions between application domains — an AI system for autonomous vehicle perception must be classified differently from one for medical image analysis, since different technical standards, safety requirements, and regulatory obligations apply to each. Second, it must acknowledge the cross-cutting character of AI: most real-world AI deployments span multiple domains simultaneously, and a taxonomy that forces exclusive domain assignment will distort rather than clarify the standards landscape. The ETSI Domain Taxonomy addresses this through multi-domain tagging, allowing any AI system or specification to receive affiliations with multiple domain nodes without hierarchy violation. This design is particularly important under the EU AI Act, which classifies AI systems by application domain for risk determination (high-risk AI systems are defined partly by the sector they operate in — Annex III to the Act lists transportation, healthcare, critical infrastructure, education, employment, and others) and where a given AI system may legitimately qualify under multiple Annex III categories. The flexibility of multi-domain tagging also reflects the practical reality that ETSI technical specifications often address multiple application contexts simultaneously: a specification on AI-driven network security (TC SAI) is simultaneously within ETSI Domain AI and ETSI Domain Security and Privacy, and must be navigable from both starting points.

    The taxonomy interfaces with wider standardisation ecosystems through careful alignment with SC 42 vocabulary (particularly ISO/IEC 22989 on AI concepts and terminology and ISO/IEC 42001 on AI management systems), with CEN-CENELEC JTC 21 harmonised standards classification, and with regulatory information systems developed under the EU AI Act. The CEN-CENELEC JTC 21 programme — which had published eight of approximately 35 planned work items by early 2026, with the first harmonised standard (prEN 18286, AI Quality Management System for EU AI Act regulatory purposes) entering public enquiry in October 2025 — uses domain classifications that must cohere with the ETSI taxonomy to avoid creating parallel, conflicting categorisation systems. ETSI participates in JTC 21 through its members and bilateral coordination arrangements, ensuring that the ETSI Domain Taxonomy’s structure serves as a compatible layer within the broader European harmonised standards programme. The ITU-T dimension is also significant: ETSI participates in ITU-T AI coordination activities (particularly through ITU-T Study Group 13, which addresses AI/ML in future networks) and ensures that ETSI domain vocabulary is compatible with ITU-T’s own AI classification schemes, reducing friction when ETSI specifications are submitted for adoption as international standards through the ITU-T A-series process.

    The practical value of the ETSI Domain Taxonomy has been demonstrated most clearly in the harmonised standards development process triggered by the EU AI Act. The European Commission’s standardisation request M/586 (May 2023), accepted by CEN-CENELEC, established the mandate for harmonised standards covering the AI Act’s Articles 9, 10, 11, 12, 13, 14, 15, and Annex IV requirements. Translating these legal requirements into specific technical standards requires first determining which application domains are at stake — a determination that uses the ETSI taxonomy as its classification vocabulary. Where ETSI’s existing work (especially TC SAI and ISG ENI output) is relevant to fulfilling the mandate, the domain taxonomy provides the traceability chain from legal requirement to application domain to technical specification. This traceability function — demonstrated through the TC SAI baseline cybersecurity standard ETSI EN 304 223 and its relationship to EU AI Act Article 15 (accuracy, robustness, and cybersecurity requirements) — represents the taxonomy’s most commercially valuable current application.

    The ETSI Domain Taxonomy is further distinguishable from adjacent classification frameworks by its institutional character: it is maintained by a recognised European Standards Organisation rather than a regulatory body, an academic consortium, or an industry association. This gives the taxonomy a form of institutional legitimacy under EU law — ETSI’s outputs carry formal presumption of conformity relevance — that frameworks from NIST, OECD, or industry consortia do not share in the EU regulatory context. The taxonomy’s classification nodes, once they underpin harmonised standards, effectively acquire quasi-legal force in the EU market access process. This institutional anchoring distinguishes the ETSI Domain Taxonomy from the ISO/IEC 42001 management system framework (which is globally applicable but does not carry EU harmonised standard status in its own right) and from the NIST AI RMF (which is a US voluntary framework with no direct EU legal relevance). The ETSI taxonomy’s European institutional grounding is its defining characteristic and the source of its practical importance in the current regulatory environment.

    Components / Architecture

    The ETSI Domain Taxonomy is structured across several interlocking architectural layers:

    • Top-Level Domain Nodes: The primary branches of the taxonomy represent broad technical-societal domains:

    • Cross-Domain Intersection Markers: ETSI’s taxonomy explicitly models cross-domain AI applications as first-class concepts:

    • Controlled Vocabulary Layer: Each domain node has canonical term definitions ensuring consistent use across ETSI technical committee outputs. Terms are aligned with ISO/IEC vocabulary where possible, reducing definitional conflicts across standards bodies.

    • Standards Mapping Layer: Domain nodes link to ETSI technical specifications, group reports, and relevant external standards, enabling navigation from application domain to applicable standardisation landscape. This layer is critical for EU AI Act compliance workflows.

    • Semantic Technology Interface: The taxonomy is designed to interface with Ontology and Semantic Web tools, making ETSI standards artefacts machine-readable. This supports automated cross-referencing between standards databases and regulatory catalogues.

    • Governance and Maintenance Layer:

      • Maintained through ETSI ISG AI and broader Technical Committee coordination

      • Update cycles tied to ETSI work programme and EU AI Act implementing acts

      • Bilateral alignment with CEN-CENELEC JTC 21 and ISO/IEC JTC 1/SC 42

        Use Cases / Major Families

    • EU AI Act Compliance Navigation: The primary commercial use case as of 2025-2026. Manufacturers of AI systems must identify applicable harmonised standards by application domain. The ETSI Domain Taxonomy provides the domain classification layer linking AI systems to the correct technical standards under the EU AI Act’s Annex III high-risk categories and related implementing acts. The JTC 21 harmonised standards programme (targeting ~35 standards) is mapped against ETSI domain classifications, so a manufacturer of an AI-based medical diagnostic tool can navigate from “healthcare AI” in the taxonomy to applicable standards for safety, accuracy, and transparency.

    • Standards Gap Analysis: ETSI Technical Committees use the taxonomy to identify application domains lacking adequate standardisation coverage. Gap analysis outputs feed directly into ETSI’s work programme, triggering new standardisation projects. CEN-CENELEC coordinates parallel gap analyses to avoid duplication and ensure coherent European coverage. The taxonomy enables systematic comparison across bodies, revealing where standards exist only from one body, or only in one regional context (e.g., NIST coverage without European equivalent).

    • Cross-Organisational Alignment: The taxonomy supports coherent presentation of ETSI outputs to SC 42, IEEE Standards, ITU-T AI Standards, and national standards bodies (BSI in UK, DIN in Germany, AFNOR in France). When ETSI domain classifications align with SC 42 vocabulary, European specifications are more readily adopted as international standards under the Vienna Agreement fast-track mechanism.

    • Machine-Readable Regulatory Infrastructure: By interfacing with Semantic Web and Knowledge Representation tools, the taxonomy feeds into machine-readable regulatory information systems. Automated cross-referencing between standards databases (ETSI’s own portal, CEN-CENELEC’s standards shop, ISO Online Browsing Platform) and regulatory catalogues (EU AI Act database, EU notified body registers) reduces compliance search overhead. This capability connects the ETSI Domain Taxonomy to broader Data Governance and Machine Learning Operations platforms that require standards traceability.

    • Research and Policy Analysis: Policy analysts at the European Commission, national ministries, and think tanks use the taxonomy to map the standardisation landscape and identify regulatory gaps. Academic researchers use it to contextualise AI system evaluations within appropriate domain standards frameworks, providing a structured vocabulary for comparative policy research. The taxonomy itself has become an object of study in the academic literature on AI governance and standardisation.

    • ETSI Work Programme Navigation: Internal ETSI use: technical committee members and industry participants use the taxonomy to determine which ETSI body has responsibility for standards in a given AI domain, preventing scope conflicts and enabling efficient referral of new work items to the appropriate group. The taxonomy provides the shared reference vocabulary that enables an expert arriving at ETSI from an AI research background (rather than a telecoms engineering background) to quickly map their domain of expertise onto the relevant ETSI work item portfolio and identify where contributions are needed.

      Standards and Context

      The ETSI Domain Taxonomy is contextualised within the following key standards frameworks and institutional relationships:

    • ETSI ISG ENI (Experiential Networked Intelligence): Established 2017. Produces specifications for AI-driven cognitive network management. Key to the ETSI Domain AI classification node. Release 4 work (2024-2026) introduces common ontologies and AI agent interface standards directly relevant to the taxonomy’s semantic layer.

    • ETSI TC SAI (Securing Artificial Intelligence): Converted from ISG SAI circa 2024. Produces cybersecurity requirements for AI systems. ETSI EN 304 223 V2.1.1 (December 2025) is the flagship output — the first global baseline AI security standard. All TC SAI outputs are classified under ETSI Domain AI ∩ ETSI Domain Security and Privacy in the taxonomy.

    • ETSI ISG MEC (Multi-access Edge Computing): Established 2014. Phase 4 (from April 2024) integrates AI/ML inference at edge nodes. Key outputs: GS MEC 003 V4.1.1 (May 2025), GR MEC 043 V4.1.1 (August 2025). All AI-related MEC specifications are classified under ETSI Domain AI in the taxonomy.

    • ETSI ISG ZSM (Zero-touch Network and Service Management): Defines AI/ML enablers for automated network operations. ZSM specifications classified under ETSI Domain AI with network management domain intersection.

    • EU AI Act (Regulation 2024/1689): The primary regulatory driver for the taxonomy’s application domain classifications. Annex III high-risk AI system categories map directly to taxonomy domain nodes. Article 40 harmonised standards presumption of conformity creates the commercial incentive for manufacturers to navigate the taxonomy.

    • CEN-CENELEC JTC 21: The European joint technical committee producing harmonised standards for EU AI Act compliance. Works with ETSI taxonomy domain classifications as the shared vocabulary for standard-to-requirement mapping. Approximately 35 harmonised standards in progress (revised completion end 2026).

    • ISO/IEC JTC 1/SC 42: The primary international AI standards committee. ISO/IEC 22989 (AI vocabulary), ISO/IEC 42001 (AI management systems), ISO/IEC 23053 (ML framework) provide the foundational vocabulary that the ETSI taxonomy aligns with. SC 42’s work feeds into the ETSI taxonomy via bilateral liaison and ETSI member participation in SC 42 plenary meetings.

    • NIST AI RMF (Risk Management Framework): A US voluntary framework providing risk management lifecycle guidance for AI systems. The ETSI taxonomy contrasts-with NIST AI RMF: where ETSI provides domain-classification-focused navigation of AI standards, NIST AI RMF provides lifecycle-focused risk management processes. Both are used together by multinational organisations operating in both EU and US markets.

    • IEEE AI Standards: IEEE 2894-2024 (XAI Architectural Framework) and related IEEE AI standards address overlapping technical territory to ETSI’s AI governance and human interface domain nodes. ETSI and IEEE maintain bilateral liaison arrangements to coordinate vocabulary and avoid conflicting standards.

    • ITU-T AI Standards: ITU-T Study Group 13 (Future Networks) and the ITU Focus Group on AI for Health (closed 2022, outputs transferred to ITU-T) produce AI standards with which ETSI coordinates to ensure compatible vocabulary. ITU-T recommendations classified under ETSI Domain AI-equivalent domains provide the international baseline against which ETSI’s European standards are positioned.

      Formal Characterisation

      The ETSI Domain Taxonomy can be formally characterised as a polyhierarchical controlled vocabulary expressed as an OWL ontology with SKOS annotations. Formally:

    • Let D = {d₁, d₂, …, dₙ} be the set of all domain nodes in the ETSI taxonomy (e.g., d₁ = ETSIDomainAI, d₂ = ETSIDomainData, d₃ = ETSIDomainSecurityAndPrivacy).

    • Let W = {w₁, w₂, …, wₘ} be the set of all ETSI work items (specifications, group reports, work packages).

    • The taxonomy defines a multi-valued classification function f: W → P(D) (the power set of D), such that each work item wᵢ is associated with a subset of domain nodes representing its application domains.

    • Cross-domain intersection nodes dᵢⱼ ∈ D are defined as nodes for which the defining characteristic is membership in two or more primary domain sets: dᵢⱼ is applicable to wₖ if and only if dᵢ ∈ f(wₖ) ∧ dⱼ ∈ f(wₖ).

    • Standards gap analysis is formalised as the identification of domain nodes dᵢ ∈ D for which the set {wₖ ∈ W : dᵢ ∈ f(wₖ)} is empty or insufficient relative to known standardisation requirements Rᵢ for that domain.

    • Harmonised standards mapping is formalised as the identification of harmonised standards hₛ applicable to an AI system aₜ with domain classification f(aₜ) ⊆ D.

      In OWL/DL terms, the taxonomy is expressed using a combination of OWL class hierarchies (SubClassOf axioms), object property domain and range restrictions (ObjectSomeValuesFrom axioms), and annotation properties (SKOS prefLabel, definition, altLabel) for vocabulary management. The taxonomy is designed to be reasoned over by standard DL reasoners (Hermit, Pellet, ELK), enabling automated classification and gap detection.

      The ETSI Domain Taxonomy can be directly compared with several adjacent AI classification and governance frameworks to clarify its distinctive role:

      vs ISO/IEC 42001 (AI Management System)

      ISO/IEC 42001:2023 is a certifiable management system standard for organisations developing, deploying, or using AI systems. Its primary classification axis is the AI system lifecycle (Plan-Do-Check-Act within an AI Management System). The ETSI Domain Taxonomy’s primary axis is application domain. These are orthogonal: a single AI system can be in the deployment phase of ISO/IEC 42001 while simultaneously being classified under multiple ETSI domain nodes. The two frameworks are increasingly used together in EU AI Act compliance programmes: ETSI taxonomy for standard identification by domain, ISO/IEC 42001 for governance lifecycle documentation. The ETSI taxonomy contrasts-with ISO/IEC 42001 in scope but is not a competitor — they address different compliance questions.

      vs NIST AI Risk Management Framework (AI RMF 1.0, 2023)

      The NIST AI RMF provides risk management lifecycle guidance for AI systems, organised around four functions: MAP (context understanding), MEASURE (risk analysis), MANAGE (risk response), and GOVERN (organisational AI governance). Like ISO/IEC 42001, NIST AI RMF is lifecycle-oriented. Unlike the ETSI taxonomy, it has no formal application domain classification structure. NIST AI RMF is a US voluntary framework with no direct EU legal relevance; the ETSI taxonomy, as output of a European Standards Organisation, carries formal relevance for EU AI Act harmonised standards. US-EU multinational organisations typically need both.

      vs EU AI Act Annex III (Risk Classification by Sector)

      The EU AI Act’s Annex III lists eight categories of high-risk AI application sectors (biometric identification, critical infrastructure, education, employment, essential services, law enforcement, migration, administration of justice and democratic processes). This list is risk-level-based (only specific uses within each sector are high-risk), while the ETSI Domain Taxonomy is coverage-based (all AI applications in a domain are within scope). The Annex III categories do not constitute a taxonomy — they are legal categories without sub-structure, controlled vocabulary, or machine-readable form. The ETSI taxonomy provides the structured, machine-readable classification vocabulary that Annex III categories lack, making it the natural bridging layer between EU AI Act requirements and the technical standards corpus.

      vs OECD AI Principles and Policy Observatory Classification

      The OECD classifies AI by technical capability type (recognition, generation, recommendation, optimisation, automation) and by application sector using a different sectoral vocabulary from ETSI’s. The OECD’s policy observatory is targeted at national governments for policy analysis rather than at manufacturers for technical standards compliance. The ETSI taxonomy’s vocabulary is aligned with ISO/IEC 22989 international standards vocabulary, while the OECD’s vocabulary reflects policy analysis needs. Cross-referencing the two vocabularies reveals which OECD application sectors map to which ETSI domain nodes.

      vs IEEE P7000 AI Ethics Standards Series

      IEEE’s AI ethics standards portfolio classifies AI concerns by ethical dimension: privacy (P7002), bias (P7003), transparency (P7001), wellbeing (P7010), autonomous system safety (P7009), algorithmic bias (P7003). These dimension-based classifications are orthogonal to ETSI’s domain-based classification. A manufacturer developing an AI system for employment screening (ETSI Domain AI in the employment sector) might need both ETSI domain-classified technical standards for the system’s capabilities and IEEE P7003 for its bias auditing procedures. The ETSI Domain AI + Governance node is the primary point of intersection between ETSI domain taxonomy and IEEE ethics standards scope.

      vs CEN-CENELEC JTC 21 Work Item Classification

      CEN-CENELEC JTC 21 classifies its approximately 35 harmonised standards work items by the EU AI Act Article or requirement they address (Article 9 risk management, Article 10 data governance, Article 13 transparency, Article 15 accuracy/robustness/cybersecurity, etc.). This legal-article-based classification is complementary to ETSI’s domain-based taxonomy: the ETSI taxonomy identifies which domain a standard applies to, while JTC 21 classification identifies which legal requirement the standard addresses. Mapping between the two classification systems is a key activity in the harmonised standards programme, ensuring complete coverage of both dimensions.

      Academic Context

      The ETSI Domain Taxonomy sits at the intersection of several scholarly traditions:

    • Taxonomy Theory and Knowledge Organisation: The foundational literature on taxonomy design (Ranganathan, 1962; Svenonius, 2000) and controlled vocabularies informs the vocabulary design choices. The shift from monohierarchical to polyhierarchical classification (Hjørland, 2017) directly justifies ETSI’s multi-domain tagging approach for AI systems. Ranganathan’s facet analysis methodology — decomposing documents/concepts into independent facets (personality, matter, energy, space, time) — has direct analogues in the ETSI taxonomy’s multi-dimensional domain classification, where an AI system specification can be simultaneously faceted by technology type, application domain, regulatory context, and deployment environment.

    • Standardisation Studies: Blind (2004) provides the economic theory of standards as coordination devices, explaining why ETSI needs a shared taxonomy even when its technical committees operate with substantial autonomy. Egyedi and Mehos (2012) address how standards bodies organise technical knowledge infrastructure. Büthe and Mattli (2011) examine the geopolitics of international standards, directly relevant to understanding ETSI’s positioning relative to ISO, IEEE, and ITU in AI standardisation. Cargill (1989) provides historical context on de jure versus de facto standardisation, framing the ETSI taxonomy as a de jure classification instrument operating in a field where de facto AI frameworks (from Google, Microsoft, OpenAI) exercise significant definitional influence.

    • AI Governance Frameworks: Dafoe (2018) and Cihon (2019) lay out research agendas for AI governance that explicitly include standardisation as a governance mechanism, positioning frameworks like the ETSI Domain Taxonomy as instruments of AI governance infrastructure. Raji et al. (2020) examine audit and accountability mechanisms for AI systems, informing the governance sub-domain taxonomy nodes that classify ETSI work on explainability and bias auditing. Jobin, Ienca, and Vayena (2019) map global AI ethics principles and their vocabulary across 84 documents, providing empirical grounding for the vocabulary choices in ETSI Domain AI + Governance classification nodes.

    • EU AI Act Scholarship: Veale and Zuiderveen Borgesius (2021) provide early analysis of the draft AI Act’s risk classification approach, identifying the centrality of application domain for risk determination — a direct argument for the importance of the ETSI Domain Taxonomy’s domain classification function. Hacker et al. (2023) examine AI regulation in the EU more broadly. The 2025 European Journal of Risk Regulation papers on technical standardisation challenges under the EU AI Act (Cambridge Core and Tandfonline) provide the most current peer-reviewed assessment of the harmonised standards situation within which the ETSI taxonomy operates.

    • Ontology Engineering: Smith et al. (2007) on the OBO Foundry principles for biomedical ontology design, and Noy and McGuinness (2001) on ontology development 101, inform the semantic technology interface design of the ETSI taxonomy. The W3C SKOS specification (Miles and Bechhofer, 2009) provides the exact vocabulary for representing controlled vocabularies as machine-readable linked data, directly applicable to the ETSI taxonomy’s semantic layer where canonical term definitions are expressed in RDF-compatible formats.

    • AI Standardisation Research: Tamburrini (2022) examines AI standardisation ethics and the question of whose values are encoded in technical standards. Ebers et al. (2021) analyse AI and digital product liability standardisation in Europe. The 2025 Skadden analysis on EU standardisation supporting the AI Act (Skadden, 2024) provides practitioner perspective on how standards bodies like ETSI are positioned in the compliance ecosystem.

    • Information Retrieval and Standards Navigation: The use of taxonomies and ontologies for navigating large corpora of technical documents (Salton and McGill, 1983; Srinivasan, 2004) provides the information retrieval theoretical basis for the ETSI taxonomy’s navigation function. As ETSI’s standards corpus exceeds several thousand documents, taxonomic navigation becomes equivalent to the information retrieval problem at scale.

      Current Landscape (2026)

      As of June 2026, the ETSI Domain Taxonomy occupies a central position in the European AI standardisation ecosystem, with its operational significance elevated substantially by the progressive enforcement of the EU AI Act:

      Regulatory Enforcement Context

      The EU AI Act entered into force August 2024. Prohibited AI practices became applicable February 2025, immediately affecting AI-based social scoring, untargeted facial recognition scraping, and emotion recognition in workplaces and schools. GPAI model obligations became effective August 2025, requiring providers of general-purpose AI models to maintain technical documentation and implement copyright policies. General provisions for high-risk AI systems under Annex III apply from August 2026, making mid-2026 the critical compliance deadline for most high-risk AI manufacturers. The ETSI Domain Taxonomy’s role in linking AI systems to applicable standards is therefore at peak operational relevance precisely at the time of writing, with manufacturers actively navigating the taxonomy to identify which harmonised standards apply to their specific AI system domain classifications.

      Harmonised Standards Programme Status

      The CEN-CENELEC JTC 21 harmonised standards programme is the primary standards delivery mechanism for EU AI Act compliance. As of early 2026, eight of approximately 35 planned work items have been published. The first harmonised standard — prEN 18286 (AI Quality Management System for EU AI Act Regulatory Purposes, aligned with ISO/IEC 42001) — entered public enquiry on 30 October 2025. The programme is running approximately eight months behind the original April 2025 completion target; the chair of CEN-CENELEC JTC 21 indicated completion was anticipated by end of 2026 at the earliest. The ETSI Domain Taxonomy provides the domain classification backbone for mapping these approximately 35 standards to specific AI system types and application contexts, making it the navigational infrastructure underpinning the entire harmonised standards architecture.

      Key ETSI Standards Milestones

      ETSI TC SAI published ETSI EN 304 223 V2.1.1 in December 2025 — the first global standard setting minimum cybersecurity requirements across the full AI lifecycle for all supply chain stakeholders. This standard, co-developed with UK DSIT and NCSC, maps threat categories against AI system domain classifications, making the ETSI taxonomy operationally essential for applying the standard correctly to specific deployment contexts. ETSI TR 104 128 V1.1.1 (May 2025) provides implementation guidance including cross-mappings to OWASP AI Exchange, MITRE ATLAS, and NIST Adversarial Attacks Taxonomy. ETSI ISG ENI’s Release 4 (GS ENI 005 V4.1.1, January 2026) establishes common ontologies enabling cross-domain automation in 6G-adjacent network environments; this ENI ontological work is technically adjacent to the ETSI Domain Taxonomy’s semantic layer, and coordination between ENI ontology outputs and the broader domain taxonomy is an active 2026 workstream.

      Institutional Developments

      The ETSI RISE board strategy group is actively recruiting academic contributors to AI standardisation, broadening the epistemic base beyond incumbent telecoms operators. ISO/IEC JTC 1/SC 42’s ongoing work on planned revisions to ISO/IEC 22989 (AI concepts and terminology) requires the ETSI taxonomy vocabulary to remain synchronised with international baseline vocabulary — an ongoing maintenance obligation that places resource demands on ETSI’s coordinating capacity but also represents a quality assurance mechanism ensuring ETSI’s domain vocabulary remains consistent with international normative reference.

      Key Terminology

      The following terms are central to understanding the ETSI Domain Taxonomy and its role in the AI standardisation ecosystem:

    • Domain Node: A named classification concept in the taxonomy representing a distinct application sector, technology area, or cross-domain intersection. Examples: ETSIDomainAI, ETSIDomainData, ETSIDomainSecurityAndPrivacy.

    • Cross-Domain Intersection Node: A taxonomy node explicitly representing the intersection of two or more primary domain concepts, enabling classification of work items that are inherently multi-domain. Examples: ETSI Domain AI + Governance, ETSI Domain AI + Data Mgmt, ETSI Domain: Data Management + AI.

    • Multi-Domain Tagging: The taxonomy design feature allowing a work item to receive multiple domain affiliations simultaneously, without hierarchy violation. Essential for capturing the cross-cutting character of AI specifications and reflecting the multi-domain applicability of most real-world AI systems.

    • Controlled Vocabulary: A standardised set of terms with canonical definitions and SKOS-compatible metadata, ensuring consistent use across ETSI technical committee outputs and alignment with ISO/IEC vocabulary (particularly ISO/IEC 22989).

    • Harmonised Standard: A European standard (EN) developed under a European Commission standardisation request by a European Standards Organisation (ETSI, CEN, or CENELEC), whose application creates a legal presumption of conformity with specified regulatory requirements in the EU single market. The ETSI Domain Taxonomy provides the domain classification navigation layer for the harmonised standards programme.

    • Standards Gap Analysis: A systematic examination of taxonomy domain nodes to identify application sectors or technology areas lacking adequate standardisation coverage. The ETSI taxonomy provides the domain node vocabulary that makes systematic gap analysis tractable across ETSI’s distributed committee structure.

    • ISG (Industry Specification Group): A flexible ETSI body open to non-ETSI-member organisations, used for faster specification development in emerging areas. Multiple ISGs (ENI, MEC, ZSM, SAI) produce AI-relevant specifications classified under the ETSI Domain Taxonomy.

    • TC (Technical Committee): A formal ETSI standards body with defined scope, producing standards with higher institutional standing than ISG deliverables. TC SAI (Securing AI) is the primary AI-security TC; TC Cyber handles broader cybersecurity including AI components.

    • EU AI Act Annex III: The list of high-risk AI application domains in the EU AI Act (biometric identification, critical infrastructure, education, employment, essential services, law enforcement, migration, justice, democratic processes), which maps onto ETSI domain taxonomy nodes, making the taxonomy a navigation tool for identifying applicable harmonised standards by legal risk classification.

    • GPAI (General Purpose AI): AI models with broad applicability across many tasks and domains simultaneously. The ETSI taxonomy’s multi-domain tagging capability is particularly relevant for GPAI classification, as GPAI systems may require classification under multiple domain nodes reflecting their diverse application contexts.

      UK Context

      The United Kingdom has multifaceted engagement with the ETSI Domain Taxonomy across government, industry, academia, and standards bodies. This engagement is particularly significant given the UK’s post-Brexit position: fully outside the EU single market but maintaining full membership of European standards bodies (ETSI as a direct member, BSI in CEN-CENELEC), creating a configuration where UK industry must navigate EU AI Act harmonised standards requirements for export market access while UK-specific AI regulation develops in parallel under the AI Safety Institute and prospective UK AI legislation.

    • BSI (British Standards Institution): As the UK national standards body and ETSI member, BSI participates in ETSI Technical Committees and ISGs that produce specifications classified under the ETSI Domain Taxonomy. BSI also represents the UK in CEN-CENELEC JTC 21, where harmonised standards mapped against ETSI taxonomy domain classifications are being produced. The post-Brexit situation means UK standardisation experts participate in European AI standards development and can shape the taxonomy’s vocabulary and classification structure, even without UK regulatory obligation to implement EU AI Act requirements domestically.

    • NCSC and DSIT: The National Cyber Security Centre and Department for Science, Innovation and Technology co-developed ETSI EN 304 223 with TC SAI, with UK AI Cyber Security Code of Practice (early 2025) anchoring UK AI security policy directly to this ETSI standard. The NCSC’s engagement makes UK government a direct shaper of content classified within the ETSI Domain AI + Governance and security nodes of the taxonomy. DSIT’s involvement signals ministerial-level recognition of the ETSI Domain Taxonomy’s strategic importance for UK AI industry market access and competitiveness.

    • BT Group Research (Adastral Park, Ipswich/Suffolk): BT chairs sessions within TC Cyber and TC SAI at ETSI, reflecting BT’s long-standing position as one of ETSI’s most active national telecom operator members. BT Research at Adastral Park has historically contributed to ETSI ISG ENI requirements, meaning BT technical work maps directly to ETSI Domain AI classification nodes in the taxonomy. BT’s AI-driven network management deployments across its national fixed and mobile infrastructure (including the EE 5G network) provide operational use case grounding for ENI specifications.

    • Vodafone UK: As a major ETSI member and pan-European operator, Vodafone contributes to ISG MEC and ISG ENI work items. Vodafone’s UK 5G network deployments across Greater London, the South East, Midlands, and Northern English cities, combined with private network AI applications for enterprise and industrial customers, are the commercial context in which ETSI Domain Taxonomy classifications are operationally applied.

    • Arm Holdings (Cambridge): Arm’s semiconductor IP underpins the vast majority of mobile and edge AI hardware globally. Arm participates in AI standardisation activities at the intersection of edge inference and telecommunications (ETSI ISG MEC), with domain content classified under ETSI Domain AI + Data Mgmt and ETSI Domain AI + Human Interface in the taxonomy. As Cambridge-headquartered UK deep tech, Arm’s standardisation work shapes how edge AI capabilities are taxonomised in ETSI.

    • The Alan Turing Institute (London): The Turing’s AI governance, fairness, and interpretability research directly informs the vocabulary and conceptual scope of ETSI Domain AI + Governance domain nodes. The Turing’s FAST track AI principles (Fairness, Accountability, Sustainability, Transparency) have influenced UK AI governance vocabulary now reflected in ETSI governance domain node definitions.

    • UK Academic Engagement:

      • Imperial College London: AI safety and robustness research groups with EU standards policy engagement; contributes to literature cited in ETSI governance work items
      • University of Edinburgh: Centre for Technomoral Futures AI ethics research informing governance taxonomy concepts; EPSRC-funded AI safety research relevant to TC SAI scope
      • University of Cambridge: Computer Laboratory AI safety research intersecting with ETSI TC SAI requirements; Judge Business School AI regulation research informing policy context
      • University of Manchester: Information School research on knowledge organisation, classification systems, and controlled vocabulary design — directly relevant to ETSI Domain Taxonomy architecture; Alan Turing Institute nodes and Jodrell Bank connected AI research
      • University of Sheffield: GATE NLP research with AI standards relevance; Information School research on information governance mapping to ETSI Domain AI + Data Mgmt

    • Northern England Industrial Context: The Manchester-Leeds-Sheffield-Newcastle-Humber industrial corridor provides the ground-level deployment context for ETSI-standardised AI systems classified under the taxonomy. Manufacturing AI (predictive maintenance at Siemens Goole, Tata Steel, and chemical processors in Teesside), healthcare AI (NHS diagnostic imaging AI across Yorkshire and Humber Integrated Care Systems, remote monitoring AI in Greater Manchester NHS trusts), transport AI (Transpennine rail optimisation, HS2 construction logistics AI, autonomous vehicle testing at MIRA in Nuneaton), and public safety AI (Greater Manchester Combined Authority AI procurement governance, Northumbria Police technology evaluation programmes) all instantiate application domains represented in the ETSI Domain Taxonomy. Smart city deployments — Sheffield Future Urban Living, Manchester CityVerve IoT/AI legacy, Newcastle Urban Observatory — map to multiple cross-domain taxonomy nodes including AI + Data Mgmt, AI + Human Interface, and AI + Governance. These real deployments give the taxonomy’s abstract domain classifications concrete industrial referents.

      Future Directions (2026-2030)

      The ETSI Domain Taxonomy faces several evolutionary pressures over the medium term, driven by technological change, regulatory development, and structural shifts in the global AI standardisation landscape:

      EU AI Act Harmonised Standards Completion and Post-Publication Phase

      The completion of the CEN-CENELEC JTC 21 harmonised standards programme (approximately 35 standards, revised completion end 2026 or early 2027) will crystallise the ETSI Domain Taxonomy’s domain classification structure as a commercially critical compliance tool. Once harmonised standards are published in the Official Journal of the European Union, they carry a legal presumption of conformity with EU AI Act requirements. This transforms the ETSI taxonomy’s domain nodes from informational navigation aids into components of a quasi-legal compliance architecture. Post-publication, the taxonomy will need a maintenance process for harmonised standard updates — when standards are revised, the domain classification must remain consistent and the compliance navigation layer must be updated to reflect new or superseded standards. ETSI’s participation in JTC 21’s ongoing maintenance regime will be essential.

      6G Native AI Taxonomy Architecture

      ETSI ISG ENI’s 6G native AI work (GR ENI 051 V4.1.1, February 2025) anticipates that in 6G networks, AI will not be an application deployed on top of network infrastructure but rather a structural component of network protocols, radio resource management, and core network functions. This “AI-native” architecture challenge — where the AI is the network, not a service running on the network — will require the ETSI Domain Taxonomy to introduce new classification nodes that do not currently exist: AI-native radio access, AI-native transport layer, AI-native core network management. These nodes will need to represent the cross-layer character of 6G native AI, breaking the current assumption that AI is an ApplicationLayer phenomenon. Coordination with 3GPP Release 19 and beyond will be necessary to ensure ETSI taxonomy nodes align with 3GPP’s own AI classification architecture for 6G.

      Agentic AI Classification Framework

      The emergence of AI agent systems — capable of goal-directed autonomous action, multi-step planning, tool use, memory, and collaboration with other agents (addressed in ETSI ISG ENI GS ENI 059 on AI Agent Interfaces, active 2025-2026) — will require new taxonomy nodes for agentic AI capabilities that do not fit cleanly within current ML inference classification nodes. Key classification challenges for agentic AI include: (1) agents that operate across multiple domain nodes simultaneously (a network management agent may use NLP for human interface, data analysis for situational awareness, and decision planning for network reconfiguration — spanning ETSI Domain AI + Human Interface, + Data Mgmt, and the core AI node simultaneously); (2) the distinction between autonomous AI agents and partially autonomous “co-pilot” AI assistants requiring different governance standards; (3) multi-agent system coordination, which may require domain nodes at the system-of-systems level rather than individual AI component level. The taxonomy’s multi-domain tagging model will need to be extended with an agency dimension that captures the degree of autonomous decision-making authority alongside the application domain classification.

      ISO/IEC 42001 Alignment and Complementarity

      As ISO/IEC 42001 (AI Management System Standard, published December 2023) gains global certification uptake — particularly among organisations seeking to demonstrate EU AI Act compliance via the management system route — the ETSI taxonomy will need to maintain tight vocabulary alignment with 42001’s lifecycle terminology (PDCA: Plan-Do-Check-Act; AI system lifecycle phases). The current contrast relationship (ETSI taxonomy is domain-classification-focused; 42001 is management-system-lifecycle-focused) is expected to evolve into a complementarity relationship, with dual-use of both frameworks becoming standard practice for EU AI Act compliance: use the ETSI taxonomy to identify applicable standards by domain, use ISO/IEC 42001 to demonstrate governance maturity. Formal vocabulary alignment mappings between ETSI taxonomy nodes and ISO/IEC 42001 risk assessment categories are an anticipated deliverable for the 2027-2028 timeframe.

      Digital Product Passport Integration

      The EU Ecodesign for Sustainable Products Regulation (ESPR, adopted 2024) introduces Digital Product Passports (DPP) — machine-readable product-level records providing supply chain traceability and sustainability information. AI systems embedded in physical products regulated under ESPR (smart home devices, industrial machinery, vehicles) will need to declare their AI capabilities in the DPP. The ETSI Domain Taxonomy, if its semantic technology interface is extended, could serve as the authoritative classification vocabulary for AI component declarations within DPPs, linking product-level AI declarations to the applicable ETSI standards corpus. This use case would transform the taxonomy from a standards navigation tool into a product lifecycle data infrastructure component.

      Global Interoperability and Non-EU Alignment

      ITU-T coordination activities, ISO participation via Vienna Agreement fast-track, and emerging non-EU AI governance frameworks (US AI Act legislative proposals at state and federal level, UK AI regulation under the AI Safety Institute and prospective AI legislation, Singapore’s Model AI Governance Framework, and Japan’s AI strategic framework) create pressure for ETSI domain taxonomy vocabulary to be globally interoperable rather than purely European in scope. A taxonomy optimised solely for EU AI Act compliance risks becoming a regional standard that impedes international market access for European AI companies. The taxonomy’s maintainers will need to make deliberate choices about which vocabulary elements are EU-specific (required for EU AI Act compliance) and which can be globally harmonised (benefiting international standard adoption via ISO/IEC JTC 1/SC 42). This global interoperability challenge is the most structurally complex medium-term evolution for the ETSI Domain Taxonomy.

      ETSI Domain Taxonomy in the Metaverse and XR Context

      This knowledge graph is primarily structured around a metaverse and extended reality (XR) ontology. The ETSI Domain Taxonomy occupies a specific structural role in this context:

      Metaverse Component Classification Function: Within the metaverse ontology, any component that is standardised or standardisable by ETSI receives a domain taxonomy classification. The ETSI Domain Taxonomy provides the ETSI-specific classification vocabulary that connects metaverse components to the applicable standards landscape. This is why the json-ld blocks in ETSI Domain AI and its sub-domain nodes reference “metaverse components” — the taxonomy nodes serve as classification anchors for metaverse AI components within ETSI’s standards framework.

      Cross-Domain Metaverse Classifications: Metaverse AI applications are inherently cross-domain. A social metaverse platform’s AI involves: Creative Media (avatar generation, environment rendering), Human Interface (natural language interaction, emotion detection), Data Mgmt (user behaviour analytics, personalisation ML pipelines), and Governance (content moderation AI, identity verification). The ETSI Domain Taxonomy’s multi-domain tagging model is specifically appropriate for this cross-domain character of metaverse AI.

      XR Standards Bridge: ETSI’s ARF (Augmented Reality Framework) work bridges the ETSI taxonomy to XR specifically. ARF specifications classified under ETSI Domain AI (for AI-driven scene understanding components) and ETSI Domain: Creative Media (for XR content) provide the standards anchor points for XR components in this knowledge graph. The ETSI GR ARF 007, ETSI GR ARF 010, ETSI GR ARF 012, ETSI ARF 010 pages in this knowledge graph represent these ARF standards artefacts.

      MEC Standards for Low-Latency Metaverse: The metaverse’s stringent latency requirements (sub-20ms for comfortable XR, sub-10ms for haptic interaction) make edge computing essential. ETSI ISG MEC specifications — classified under ETSI Domain AI in the taxonomy — define the edge computing infrastructure that enables AI inference close to the user for real-time XR rendering and interaction. The taxonomy thus connects metaverse architecture requirements to the applicable edge AI standards.

      Legacy Term ID Reference: The ETSI Domain Taxonomy entries in this knowledge graph include legacy-term-id values (e.g., 20330 for ETSI Domain AI, 20333 for ETSI Domain AI + Governance) derived from the original metaverse component identification system used before the ontology migration. These IDs trace back to the classification of metaverse components within ETSI’s standards portfolio, demonstrating that the ETSI Domain Taxonomy nodes in this graph are not abstract concepts but rather specific component classifications from a real metaverse standardisation classification exercise.

      Relationship to Adjacent Concepts in This Knowledge Graph

      The following pages in this knowledge graph relate directly to the ETSI Domain Taxonomy and should be understood in relation to it:

    • ETSI Domain AI: The primary AI domain node that is a direct hasPart of the ETSI Domain Taxonomy. ETSI Domain AI is the root classification concept for all AI-related ETSI specifications, grouping under itself four cross-domain intersection nodes: Creative Media, Data Mgmt, Governance, and Human Interface.

    • ETSI Domain AI + Creative Media: Cross-domain intersection node (AI × Creative Media) that is a child of ETSI Domain AI within the taxonomy. Represents the subset of ETSI work items addressing AI applied to creative content, media production, generative systems, and immersive entertainment.

    • ETSI Domain AI + Data Mgmt: Cross-domain intersection node (AI × Data Management) representing ETSI work items on ML pipeline management, intelligent data processing, and AI-driven data governance. Bridges to InfrastructureDomain and ETSI Domain Data.

    • ETSI Domain AI + Governance: Cross-domain intersection node (AI × Governance) representing ETSI work items on responsible AI, explainability, bias auditing, and regulatory compliance. Bridges to TrustAndGovernanceDomain and AI Governance Framework.

    • ETSI Domain AI + Human Interface: Cross-domain intersection node (AI × Human Interface) representing ETSI work items on conversational AI, gesture recognition, emotion detection, and adaptive user experience. Bridges to InteractionDomain and Telecollaboration.

    • ETSI Domain Data: Sibling domain node to ETSI Domain AI in the taxonomy. Represents data management, analytics, and data governance as the primary domain (as opposed to AI). Cross-domain intersection nodes bridge between ETSI Domain Data and ETSI Domain AI: ETSI Domain: Data Management + AI, ETSI Domain: Data Management + Creative Media, ETSI Domain: Data Management + Ethics, ETSI Domain: Data Management + Security.

    • ETSI Domain Security and Privacy: Sibling domain node. Represents cybersecurity and privacy as primary domains. ETSI TC SAI output is classified at the intersection of ETSI Domain AI and ETSI Domain Security and Privacy, making this sibling node the primary cross-reference point for AI security specifications.

    • ETSI Domain Identity and Trust: Sibling domain node. Digital identity, trust frameworks, and verifiable credentials as primary domain. AI-driven identity verification and authentication cross-reference this sibling node.

    • ETSI Domain: Governance & Compliance and ETSI Domain: Governance & Ethics: Governance-focused domain nodes that cross-reference with ETSI Domain AI + Governance to represent the governance domain’s perspective on AI.

    • ETSI Domain: Ethics & Law: Ethics and legal domain node providing the legal context within which AI governance specifications are positioned. Directly relevant to ETSI Domain AI + Governance and the EU AI Act harmonised standards programme.

    • ETSI ARF 010, ETSI GR ARF 007, ETSI GR ARF 010, ETSI GR ARF 012: Augmented Reality Framework specifications from ETSI, classified at the intersection of ETSI Domain AI (for AI-driven scene understanding) and other ETSI domains. These pages represent concrete standards artefacts whose classification in the taxonomy is directly relevant.

    • ETSI ENI 008: An ENI specification artefact classified under ETSI Domain AI in the taxonomy, representing the network intelligence domain’s AI management framework.

    • ETSI GR MEC 032: An MEC Group Report classified under ETSI Domain AI, addressing AI/ML service applications in the edge computing context.

    • AI Governance Framework, AI Governance, AI Governance Law and Privacy: Knowledge graph nodes representing the governance framework family to which ETSI Domain AI + Governance connects via the taxonomy’s governance dimension.

    • Machine Learning Operations: MLOps concept that bridges-to from ETSI Domain Taxonomy, representing the operational deployment context where ETSI-classified AI standards are applied in production environments.

    • Data Governance: Data governance concept that bridges-to from ETSI Domain Taxonomy, representing the data management governance context where ETSI Domain AI + Data Mgmt specifications apply.

    • EU AI Act: The regulatory instrument that has made the ETSI Domain Taxonomy commercially critical, driving the harmonised standards programme that the taxonomy navigates.

    • CEN-CENELEC: The European Standards Organisation sibling to ETSI; JTC 21 is the joint body producing harmonised standards for EU AI Act compliance, using ETSI domain classifications as the navigational vocabulary.

    • SC 42: The primary international AI standards committee; ETSI Domain Taxonomy vocabulary aligns with SC 42 output (ISO/IEC 22989, 42001) to enable global interoperability.

      Research & Literature

      1. ETSI (2024-2025). Work Programme 2024-2025: Advancing Standards. ETSI. https://www.etsi.org/images/files/WorkProgramme/ETSI-Work-Programme-2024-2025.pdf — Official ETSI biennial work programme listing all active work items and their domain classifications; primary source for understanding the ETSI Domain Taxonomy’s current coverage.
      2. ETSI White Paper No. 52. ETSI Activities in the Field of Artificial Intelligence. ETSI. https://www.etsi.org/images/files/ETSIWhitePapers/ETSI-WP52-ETSI-activities-in-the-field-of-AI.pdf — Overview of ETSI’s distributed AI standardisation approach; documents the 14+ Technical Committees and ISGs producing AI-relevant specifications classified in the taxonomy.
      3. ETSI EN 304 223 V2.1.1 (2025-12). Securing Artificial Intelligence (SAI); Baseline Cyber Security Requirements for AI Models and Systems. ETSI — The flagship TC SAI output; first European and first global baseline AI security standard; classified under ETSI Domain AI ∩ Security and Privacy in the taxonomy.
      4. ETSI TR 104 128 V1.1.1 (2025-05). Securing Artificial Intelligence (SAI). ETSI. https://www.etsi.org/deliver/etsi_tr/104100_104199/104128/01.01.01_60/tr_104128v010101p.pdf — Implementation guidance for ETSI EN 304 223 with cross-mappings to OWASP, MITRE ATLAS, and NIST Adversarial Attacks Taxonomy.
      5. ETSI TR 104 065 V1.1.1 (2025-05). Securing Artificial Intelligence (SAI). ETSI. https://www.etsi.org/deliver/etsi_tr/104000_104099/104065/01.01.01_60/tr_104065v010101p.pdf — Technical Report providing context and analysis for TC SAI cybersecurity requirements.
      6. ETSI GS ENI 005 V4.1.1 (2026-01). Experiential Networked Intelligence (ENI); ENI System Architecture. ETSI. https://www.etsi.org/deliver/etsi_gs/ENI/001_099/005/04.01.01_60/gs_ENI005v040101p.pdf — Release 4 ENI system architecture; introduces AI agent interfaces and common ontologies for cross-domain automation.
      7. ETSI GR ENI 051 V4.1.1 (2025-02). Experiential Networked Intelligence (ENI); Analysis and Study of 6G Native AI Capabilities. ETSI. https://www.etsi.org/deliver/etsi_gr/ENI/001_099/051/04.01.01_60/gr_ENI051v040101p.pdf — Forward-looking Group Report on 6G native AI; key document for understanding forthcoming ETSI Domain AI taxonomy extensions.
      8. ETSI GS MEC 003 V4.1.1 (2025-05). Multi-access Edge Computing (MEC); Framework and Reference Architecture. ETSI. https://www.etsi.org/deliver/etsi_gs/mec/001_099/003/04.01.01_60/gs_mec003v040101p.pdf — Phase 4 MEC reference architecture; primary classification anchor for edge AI specifications in ETSI Domain AI.
      9. European Commission (2024). Regulation (EU) 2024/1689 (EU AI Act). Official Journal of the European Union — Primary regulatory driver for the ETSI Domain Taxonomy’s commercial importance; Annex III application domain categories map to taxonomy nodes.
      10. European Commission (2024). Standardisation of the AI Act. Digital Strategy. https://digital-strategy.ec.europa.eu/en/policies/ai-act-standardisation — European Commission’s standardisation strategy page; context for standardisation request M/586 accepted by CEN-CENELEC.
      11. European Commission (2024). Harmonised Standards for the European AI Act. AI Watch. https://ai-watch.ec.europa.eu/news/harmonised-standards-european-ai-act-2024-10-25_en — October 2024 update on harmonised standards programme status.
      12. CEN-CENELEC (2025). JTC 21 Work Programme Dashboard. CEN-CENELEC — Living dashboard of approximately 35 harmonised standards work items for EU AI Act compliance; the programme that the ETSI Domain Taxonomy is designed to navigate.
      13. Standards and the EU AI Act: Legitimacy, State of Play, and Future Challenges (2025). European Journal of Risk Regulation / Taylor & Francis. https://www.tandfonline.com/doi/full/10.1080/13600834.2025.2570966 — Peer-reviewed analysis of the legitimacy and delivery challenges of the harmonised standards programme; confirms eight-month delay relative to original April 2025 deadline.
      14. European AI Standards — Technical Standardisation and Implementation Challenges under the EU AI Act (2025). Cambridge Core / European Journal of Risk Regulation. https://www.cambridge.org/core/journals/european-journal-of-risk-regulation/article/european-ai-standards-technical-standardisation-and-implementation-challenges-under-the-eu-ai-act/E5157BA0391FFA9E1A3233E636005192 — Academic analysis of implementation challenges; key reference for understanding the taxonomy’s role in the harmonised standards architecture.
      15. EU Standardization Supporting the Artificial Intelligence Act (2024). Skadden Insights. https://www.skadden.com/insights/publications/2024/10/eu-standardization-supporting-the-artificial-intelligence-act — Practitioner analysis of how ETSI, CEN, and CENELEC are positioned in EU AI Act compliance; useful for understanding the taxonomy’s market context.
      16. ISO/IEC 42001:2023. Information Technology — Artificial Intelligence — Management System. ISO/IEC JTC 1/SC 42 — World’s first certifiable AI management system standard; contrasts-with ETSI Domain Taxonomy in being lifecycle-focused; increasingly used alongside the taxonomy for EU AI Act compliance.
      17. ISO/IEC 22989:2022. Information Technology — Artificial Intelligence — Concepts and Terminology. ISO/IEC JTC 1/SC 42 — Foundation vocabulary standard; the ETSI Domain Taxonomy’s controlled vocabulary aligns with this standard to ensure international compatibility.
      18. NCSC (2025). New ETSI Standard Protects AI Systems from Evolving Cyber Threats. https://www.ncsc.gov.uk/blog-post/new-etsi-standard-protects-ai-systems-from-evolving-cyber-threats — UK government endorsement of ETSI EN 304 223; demonstrates UK-ETSI collaboration on standards classified within the ETSI Domain AI taxonomy.
      19. NCSC Annual Review (2025). Artificial Intelligence Chapter. https://www.ncsc.gov.uk/collection/ncsc-annual-review-2025/ — UK national AI security strategy context; documents NCSC’s TC SAI engagement.
      20. UK DSIT (2025). AI Cyber Security Code of Practice. Department for Science, Innovation and Technology — UK regulatory instrument built on ETSI EN 304 223; demonstrates real-world deployment of ETSI-taxonomy-classified AI security standards in national policy.
      21. Dafoe, A. (2018). AI Governance: A Research Agenda. Future of Humanity Institute, University of Oxford — Research programme positioning standardisation as a core AI governance mechanism; intellectual context for the ETSI Domain Taxonomy’s governance function.
      22. Cihon, P. (2019). Standards for AI Governance: International Standards to Enable Global Coordination in AI Research and Development. Future of Humanity Institute — Analysis of international AI standards landscape; directly relevant to ETSI’s role and the taxonomy’s global interoperability challenge.
      23. Blind, K. (2004). The Economics of Standards: Theory, Evidence, Policy. Edward Elgar — Economic theory of standards as coordination devices; explains why ETSI needs a shared domain taxonomy across its distributed committee structure.
      24. Büthe, T. & Mattli, W. (2011). The New Global Rulers: The Privatization of Regulation in the World Economy. Princeton University Press — Geopolitics of international standards; context for ETSI-ISO-IEEE positioning in AI standardisation.
      25. Jobin, A., Ienca, M. & Vayena, E. (2019). The Global Landscape of AI Ethics Guidelines. Nature Machine Intelligence, 1, 389-399 — Maps AI ethics principles across 84 global documents; empirical basis for ETSI Domain AI + Governance vocabulary choices.
      26. Noy, N.F. & McGuinness, D.L. (2001). Ontology Development 101: A Guide to Creating Your First Ontology. Stanford Knowledge Systems Laboratory — Foundational ontology engineering methodology; directly applicable to the ETSI taxonomy’s OWL/SKOS implementation.
      27. Miles, A. & Bechhofer, S. (2009). SKOS Simple Knowledge Organization System Reference. W3C Recommendation — Specification for machine-readable controlled vocabularies; directly applicable to the ETSI taxonomy’s semantic technology interface.
      28. Veale, M. & Zuiderveen Borgesius, F. (2021). Demystifying the Draft EU Artificial Intelligence Act. Computer Law Review International, 22(4), 97-112 — Early AI Act scholarship identifying domain-based risk classification as the Act’s central mechanism; direct argument for the importance of the ETSI Domain Taxonomy.

      Governance and Maintenance of the Taxonomy

      The ETSI Domain Taxonomy is a living instrument subject to ongoing maintenance as technology evolves, new standardisation work begins, and regulatory requirements change. The governance structure is as follows:

      Maintenance Responsibility

      Primary governance responsibility rests with the ETSI Board Strategy Group on Research, Innovation and Standardisation Ecosystem (RISE), which sets ETSI’s overall AI standardisation strategy and owns the taxonomy’s top-level structure. Individual Technical Committees and Industry Specification Groups bear responsibility for maintaining the classified taxonomy nodes relevant to their scope: ISG ENI owns the taxonomy nodes covering cognitive network AI, TC SAI owns the AI security domain nodes, ISG MEC owns the edge computing AI nodes, and ISG ZSM owns the network automation AI nodes.

      Update Triggers

      The taxonomy is updated in response to:

    • New ETSI work items being approved: every new work item entering ETSI’s work programme must be assigned at least one taxonomy domain classification, triggering taxonomy extensions if no existing node is appropriate

    • Regulatory developments: EU AI Act implementing acts, delegated acts, and revised Annex III lists create pressure to align taxonomy domain nodes with updated legal categories

    • Major technology shifts: the emergence of agentic AI, 6G native AI, and federated learning paradigms each create new domain classification requirements

    • ISO/IEC vocabulary updates: when ISO/IEC JTC 1/SC 42 revises foundational vocabulary (ISO/IEC 22989 is periodically revised), ETSI’s controlled vocabulary layer must be updated to maintain alignment

      Alignment with CEN-CENELEC JTC 21

      A particularly significant maintenance obligation is maintaining coherence with CEN-CENELEC JTC 21’s harmonised standards classification. JTC 21 uses a requirements-based classification (by EU AI Act Article) while ETSI uses domain-based classification; keeping these coherent requires ongoing bilateral mapping activity. A joint ETSI-JTC 21 coordination group manages this alignment, with meetings at least quarterly during the active harmonised standards production period (2024-2026).

      Version Control and Change Management

      Changes to taxonomy domain node definitions, scope descriptions, and controlled vocabulary terms are subject to ETSI’s standard change management process, which requires consensus among affected technical committee members. Significant structural changes (adding or removing top-level domain nodes, redefining cross-domain intersection criteria) require Board Strategy Group approval. Minor vocabulary additions and clarifications can be managed within individual ISG/TC processes.

      Long-Term Maintenance Commitment

      Unlike ad hoc classification schemes, the ETSI Domain Taxonomy carries an implicit long-term maintenance commitment: because manufacturers are building compliance workflows and standards navigation tools on the taxonomy’s domain node structure, changes to that structure have commercial consequences. ETSI has committed through its work programme process to maintaining stable, consistent taxonomy domain classifications for the life of the harmonised standards that depend on them (at minimum through the EU AI Act’s mandatory review period, currently anticipated 2028-2030).

      Domain Coverage: Cross-Domain Intersection Nodes in the AI Branch

      The following pages in this knowledge graph are cross-domain intersection nodes that sit within the AI branch of the ETSI Domain Taxonomy:

      Primary AI Sub-Domain Nodes (direct hasPart of ETSI Domain AI)

    • ETSI Domain AI + Creative Media: AI × Creative Media intersection. Scope: generative AI in media, deepfake detection, AI-assisted content creation, ARF AI scene understanding, immersive XR AI components.

    • ETSI Domain AI + Data Mgmt: AI × Data Management intersection. Scope: ML pipelines, AI data lakes, intelligent analytics, AI data plane at edge, AI-driven data governance.

    • ETSI Domain AI + Governance: AI × Governance intersection. Scope: responsible AI, explainability, bias auditing, EU AI Act technical requirements, AI cybersecurity requirements.

    • ETSI Domain AI + Human Interface: AI × Human Interface intersection. Scope: conversational AI, voice assistants, gesture/emotion recognition, adaptive UX, AI-driven accessibility.

      AI as Secondary Dimension (Data Management Primary Domain)

    • ETSI Domain: Data Management + AI: Data Management primary, AI as enabling technology. Focuses on data governance requirements imposed by AI system data dependencies.

    • ETSI Domain: Data Management + Creative Media: Data governance for creative media production and distribution systems.

    • ETSI Domain: Data Management + Cultural Heritage: Data governance for cultural heritage digitisation and AI analysis.

    • ETSI Domain: Data Management + Ethics: Ethical dimensions of AI data management, training data curation, and bias in data pipelines.

    • ETSI Domain: Data Management + Security: Security requirements for AI training data and inference data pipelines.

      Governance-Primary Domain Intersections

    • ETSI Domain: Governance & Compliance: Regulatory compliance frameworks for AI systems, including EU AI Act conformity assessment.

    • ETSI Domain: Governance & Ethics: Ethical governance frameworks for AI development and deployment.

    • ETSI Domain: Ethics & Law: Legal context for AI governance, liability, and regulatory compliance.

    • ETSI Domain: Creative Media: Creative media domain without AI-specific focus; context node for ETSI Domain AI + Creative Media.

      Summary of AI Domain Classification Nodes Documented in This Knowledge Graph

      The following is a complete inventory of ETSI Domain Taxonomy AI classification nodes documented within this knowledge graph, providing a navigational overview of the AI domain coverage:

    Domain NodeTypePrimary FocusKey ETSI Group
    ETSI Domain AIRoot Domain NodeAll AI capabilitiesENI, TC SAI, MEC, ZSM
    ETSI Domain AI + Creative MediaCross-Domain IntersectionGenerative AI, XR, mediaARF, TC SAI (deepfakes)
    ETSI Domain AI + Data MgmtCross-Domain IntersectionML pipelines, analyticsENI, MEC, ZSM
    ETSI Domain AI + GovernanceCross-Domain IntersectionEthics, compliance, securityTC SAI, TC Cyber
    ETSI Domain AI + Human InterfaceCross-Domain IntersectionConversational AI, UXHuman Factors, ARF
    ETSI Domain: Data Management + AIReciprocal IntersectionData governance for AIENI data specs
    ETSI Domain: Governance & ComplianceGovernance DomainRegulatory complianceTC SAI governance
    ETSI Domain: Governance & EthicsGovernance DomainAI ethics frameworksETSI AI ethics work
    ETSI Domain: Ethics & LawLegal DomainAI liability, regulationLegal/regulatory context
    ETSI Domain: Data Management + EthicsData-Ethics IntersectionEthical data governanceAI training data ethics
    ETSI Domain: Data Management + SecurityData-Security IntersectionSecure AI data pipelinesTC SAI, TC Cyber

    This inventory reflects the taxonomy’s coverage of AI domain intersections as of June 2026. Additional nodes for 6G native AI, agentic AI systems, and federated learning are anticipated in the 2027-2028 timeframe.

    Key Formal Properties and Identifiers

    The following formal properties identify and characterise the ETSI Domain Taxonomy as an ontological entity in this knowledge graph and in the broader ETSI taxonomy infrastructure:

    • OWL IRI: urn:ngm:class:etsi-domain-taxonomy
    • VisionFlow Page IRI: urn:visionflow:page:8a4313c0d8a7a59ac4839449416255bc5dd6bdf149014fcaee2244361c1ec17e
    • Alias: ETSI AI Domain Classification (sameAs: urn:ngm:class:etsi-ai-domain-classification)
    • Quality Score: 0.72 (enriched, up from lower initial quality)
    • Inference Rule: ManualEnrichment (indicating human-curated rather than automatically inferred content)
    • Domain: artificial-intelligence
    • Maturity: emerging (reflecting ongoing development of the harmonised standards programme)
    • Super-class: AI Governance and Ethics (urn:ngm:class:ai-governance-and-ethics)
    • Same-as relations: urn:ngm:class:etsi-ai-domain-classification
    • ETSI AI Technologies Page: https://www.etsi.org/technologies/artificial-intelligence-automation/
    • ETSI AI Conference: https://www.etsi.org/newsroom/press-releases/2325-etsi-artificial-intelligence-conference
    • Related external vocabulary: ISO/IEC 22989:2022 (AI concepts and terminology) — canonical alignment target for ETSI taxonomy controlled vocabulary
    • Standards alignment: CEN-CENELEC JTC 21 harmonised standards programme, ISO/IEC JTC 1/SC 42 international AI standards
    • Regulatory alignment: EU AI Act (Regulation 2024/1689), UK AI Cyber Security Code of Practice (2025)
    • Semantic technology interface: OWL DL, SKOS, RDF, SPARQL
    • Classification model: Polyhierarchical multi-domain tagging with cross-domain intersection nodes
    • Update frequency: Living instrument — updated with each new ETSI work programme cycle (biennial) and on an ad hoc basis for regulatory developments
    • Governed by: ETSI RISE Board Strategy Group, individual ISG/TC scope decisions, ETSI-CEN-CENELEC bilateral alignment
    • Key partner bodies: CEN-CENELEC JTC 21, ISO/IEC JTC 1/SC 42, ITU-T Study Group 13, BSI (UK), DIN (Germany), AFNOR (France)
    • Key UK stakeholders: NCSC, DSIT, BT, Vodafone, Arm, Alan Turing Institute, BSI
    • Key sub-domain pages in this graph: ETSI Domain AI, ETSI Domain Data, ETSI Domain Security and Privacy, ETSI Domain Identity and Trust, ETSI Domain AI + Creative Media, ETSI Domain AI + Data Mgmt, ETSI Domain AI + Governance, ETSI Domain AI + Human Interface
    • Key standards artefact pages in this graph: ETSI ARF 010, ETSI GR ARF 007, ETSI GR ARF 010, ETSI GR ARF 012, ETSI ENI 008, ETSI GR MEC 032

    Provenance

    Graph View

    Backlinks