@prefix : . @prefix rdf: . @prefix rdfs: . @prefix owl: . @prefix xsd: . @prefix skos: . @prefix schema: . @prefix dct: . @prefix foaf: . @prefix prov: . : a owl:Ontology ; schema:name "Libraries as AI Infrastructure: The African Edition — KG Ontology"@en ; schema:description "Semantic model of Stuart Michael Edelenbos's analysis of African GLAM institutions building linked data and AI infrastructure — from ALMEDA's oral literature ontology to Nigeria's SNAR authority control and the Dar Gnawa Museum's data sovereignty."@en ; rdfs:label "Libraries as AI Infrastructure: The African Edition — Knowledge Graph" ; schema:identifier "https://www.linkedin.com/pulse/libraries-ai-infrastructure-african-edition-stuart-michael-edelenbos-bsage/" ; dct:source ; dct:creator :stuartEdelenbos ; dct:date "2026"^^xsd:date . :AfricanInitiative a owl:Class ; rdfs:label "African GLAM Initiative" ; rdfs:comment "A linked data, AI, or authority control project led by African institutions to build semantic infrastructure on their own terms." ; rdfs:isDefinedBy : . :KeyConcept a owl:Class ; rdfs:label "Key Concept" ; rdfs:comment "A structural insight about knowledge infrastructure, data sovereignty, or the relationship between linked data and AI in the African context." ; rdfs:isDefinedBy : . :article a schema:Article, schema:SocialMediaPosting ; rdfs:label "Libraries as AI Infrastructure: The African Edition" ; schema:name "Libraries as AI Infrastructure: The African Edition"@en ; schema:headline "Libraries as AI Infrastructure: The African Edition"@en ; schema:description "The 'libraries as AI infrastructure' story isn't just a European story. African GLAM institutions are building linked data, authority control, and AI systems calibrated to their own knowledge traditions — not waiting for permission from European standards bodies."@en ; schema:url ; schema:author :stuartEdelenbos ; schema:publisher :linkedIn ; schema:about :linkedDataInfrastructure, :dataSovereignty, :africanGLAM, :authorityControl ; schema:hasPart :blindSpotsSection, :knowledgeHeritageSection, :almedaSection, :narSection, :darGnawaSection, :aiIntegrationSection, :reflectionSection, :faqPage, :glossarySet, :howtoSection, :initiativesSection, :conceptsSection ; schema:isPartOf :linkedInPublicationIssue ; prov:wasGeneratedBy :kgGeneratorSkill . :linkedIn a schema:Organization ; rdfs:label "LinkedIn" ; schema:url . :linkedInPublicationIssue a schema:PublicationIssue ; rdfs:label "LinkedIn Article — 2026" ; schema:hasPart :article . :stuartEdelenbos a foaf:Person, schema:Person ; schema:name "Stuart Michael Edelenbos"@en ; schema:url ; schema:jobTitle "Product Analyst @ OCLC | Syndeo"@en ; schema:affiliation :oclc ; rdfs:comment "Product Analyst at OCLC, working on WorldCat Entities. Writes about libraries as AI infrastructure with a focus on linked data, authority control, and knowledge sovereignty."@en . :oclc a schema:Organization ; rdfs:label "OCLC" ; schema:url ; rdfs:comment "Global library cooperative operating WorldCat and VIAF — the Virtual International Authority File with 60 million authority records."@en . :linkedDataInfrastructure a :KeyConcept, skos:Concept, schema:DefinedTerm ; schema:name "Linked Data as AI Infrastructure"@en ; schema:description "The argument that well-structured, semantically grounded, authority-controlled library metadata is the foundation AI systems need for trustworthy reasoning. Libraries built this infrastructure long before AI needed it — now it is essential for entity disambiguation and knowledge graph quality."@en ; rdfs:isDefinedBy : . :dataSovereignty a :KeyConcept, skos:Concept, schema:DefinedTerm ; schema:name "Data Sovereignty"@en ; schema:description "The principle that data about a community's cultural heritage should be governed by that community — stored in its own infrastructure, under its own legal frameworks. 'Clouded coloniality' describes what happens when heritage data is processed in external cloud infrastructure under foreign legal frameworks."@en ; rdfs:isDefinedBy : . :africanGLAM a :KeyConcept, skos:Concept, schema:DefinedTerm ; schema:name "African GLAM Institutions"@en ; schema:description "Galleries, Libraries, Archives, and Museums in Africa building linked data, authority control, and AI infrastructure calibrated to their own knowledge traditions — not importing wholesale from Anglo-Saxon or European models."@en ; rdfs:isDefinedBy : . :authorityControl a :KeyConcept, skos:Concept, schema:DefinedTerm ; schema:name "Semantic Authority Control"@en ; schema:description "The process of disambiguating entities (persons, works, organizations) in library metadata — turning strings into connected entities in a linked graph. VIAF holds 60M records but includes only 3 African institutions out of 55 countries."@en ; rdfs:isDefinedBy : . :initiativesSection a schema:ArticleSection ; schema:name "African GLAM Initiatives"@en ; schema:position 7 ; schema:hasPart :almedaProject, :snarProject, :darGnawaProject, :annifAfrica . :almedaProject a :AfricanInitiative ; rdfs:label "ALMEDA — African Literary Metadata Project" ; rdfs:comment "Led by Ashleigh Harris (Uppsala) with Kenyatta University and University of Lagos. Builds a multilingual metadata ontology for African genres in African languages via Wikibase — from the epistemological ground up rather than within European frameworks. Addresses the question: what if the ontology itself is wrong for oral and informal literatures?"@en . :snarProject a :AfricanInitiative ; rdfs:label "SNAR — Semantic Name Authority Repository (National Library of Nigeria)" ; rdfs:comment "A Wikibase-based semantic authority control system for Nigerian scholarly and cultural works. Links library records to Wikipedia and Wikidata, making Nigerian contributions visible to global discovery systems. Addresses VIAF's thin coverage: only 3 African institutions out of 55 countries."@en . :darGnawaProject a :AfricanInitiative ; rdfs:label "Kouyou — Dar Gnawa Museum AI Tool (Marrakech)" ; rdfs:comment "Independently built AI tool storing all data in certified national data centers in Morocco — a deliberate data sovereignty choice. Rejects external cloud processing for heritage data, applying hard-won historical lessons about extraction dynamics."@en . :annifAfrica a :AfricanInitiative ; rdfs:label "AI-Assisted Subject Indexing for African Libraries" ; rdfs:comment "Lightweight AI implementations using open-source Python and Hugging Face models, trained on African datasets, designed for resource constraints. Query response times reduced from days to milliseconds."@en . :conceptsSection a schema:ArticleSection ; schema:name "Key Insights"@en ; schema:position 8 ; schema:hasPart :vindicationVsConstruction, :cloudedColoniality, :ontologyAsSovereignty, :eurocentricMetadata . :vindicationVsConstruction a :KeyConcept ; rdfs:label "Vindication vs. Construction Narrative" ; rdfs:comment "European linked data is a vindication narrative: look what libraries built before anyone knew it was valuable. African linked data is a construction narrative: building the foundation and AI applications simultaneously, with purpose visible from the start."@en . :cloudedColoniality a :KeyConcept ; rdfs:label "Clouded Coloniality" ; rdfs:comment "What happens when heritage data is processed in external cloud infrastructure under foreign legal frameworks — the extraction dynamic of colonialism reappearing in digital form through structural arrangements convenient to the more powerful party."@en . :ontologyAsSovereignty a :KeyConcept ; rdfs:label "Ontology Design as Cultural Sovereignty" ; rdfs:comment "Building metadata frameworks from the epistemological ground up for knowledge forms that European ontologies distort — oral literature, performance traditions, collective authorship. Not finding workarounds in existing frameworks but building frameworks calibrated to the knowledge being described."@en . :eurocentricMetadata a :KeyConcept ; rdfs:label "Eurocentric Metadata Bias" ; rdfs:comment "Global discovery metadata frameworks built around European assumptions about 'work,' 'authorship,' and legitimate knowledge forms. MARC records and BIBFRAME were not designed for oral literature, performance traditions, or collective authorship — making vast cultural heritage systematically invisible to AI."@en . :blindSpotsSection a schema:ArticleSection ; schema:name "A Confession About Blind Spots"@en ; schema:position 1 ; schema:description "The author acknowledges his Eurocentric professional map — Sweden's Libris, Finland's Finto, the KB in The Hague, the British Library. His mental defaults obscured the linked data and AI work happening across African GLAM institutions right now."@en . :knowledgeHeritageSection a schema:ArticleSection ; schema:name "The Systematic Invisibility of African Knowledge"@en ; schema:position 2 ; schema:description "Much of Africa's knowledge heritage is invisible to global discovery systems — not because it doesn't exist, but because metadata frameworks were built around European assumptions. MARC records weren't designed for Gnawa ritual music, Swahili oratory, or Zimbabwe's performance traditions."@en . :almedaSection a schema:ArticleSection ; schema:name "ALMEDA: What If the Ontology Itself Is Wrong?"@en ; schema:position 3 ; schema:description "Standard ontologies organize around 'work,' 'author,' 'text,' 'publication' — concepts that don't hold for African oral and informal literatures. ALMEDA builds a multilingual metadata ontology from the epistemological ground up via Wikibase."@en . :narSection a schema:ArticleSection ; schema:name "SNAR: Entity Disambiguation at Scale"@en ; schema:position 4 ; schema:description "The National Library of Nigeria's Wikibase project builds semantic authority control for Nigerian works — addressing VIAF's thin coverage of African scholars by creating the same disambiguation infrastructure that makes VIAF valuable for European entities."@en . :darGnawaSection a schema:ArticleSection ; schema:name "Dar Gnawa: Data Sovereignty in Practice"@en ; schema:position 5 ; schema:description "The Dar Gnawa Museum built Kouyou independently, storing heritage data in Moroccan national data centers. A deliberate rejection of external cloud processing — applying historical lessons about extraction to AI infrastructure."@en . :aiIntegrationSection a schema:ArticleSection ; schema:name "AI Integration Is Already Happening"@en ; schema:position 6 ; schema:description "AI-assisted subject indexing tools for African academic libraries, trained on African datasets. Lightweight, open-source, dramatically faster than traditional cataloging. The foundation is being built in the same generation as the AI applications it will support."@en . :reflectionSection a schema:ArticleSection ; schema:name "What This Changes About the Argument"@en ; schema:position 7 ; schema:description "The European version is a vindication narrative. The African version is a construction narrative — purpose visible from the start. Some of the most interesting chapters of the libraries-as-AI-infrastructure story are being written in places the author's Eurocentric map barely marked."@en . :faqPage a schema:FAQPage ; schema:name "Frequently Asked Questions"@en ; schema:mainEntity :faq1, :faq2, :faq3, :faq4, :faq5, :faq6, :faq7, :faq8, :faq9, :faq10, :faq11, :faq12 . :faq1 a schema:Question ; schema:name "What is the 'libraries as AI infrastructure' argument?"@en ; schema:acceptedAnswer :a1 . :a1 a schema:Answer ; schema:text "Well-structured, semantically grounded, authority-controlled library metadata is the foundation AI systems need for trustworthy reasoning. Libraries built this infrastructure before AI existed — now entity disambiguation, linked data, and knowledge graphs are essential for AI that doesn't hallucinate."@en . :faq2 a schema:Question ; schema:name "What is ALMEDA and why does it matter?"@en ; schema:acceptedAnswer :a2 . :a2 a schema:Answer ; schema:text "The African Literary Metadata project, led by Ashleigh Harris with Kenyatta University and University of Lagos, builds a multilingual metadata ontology for African genres. It starts from the premise that standard bibliographic ontologies built around 'work,' 'author,' and 'text' don't fit oral and informal literatures — and builds a Wikibase-based framework from the epistemological ground up."@en . :faq3 a schema:Question ; schema:name "What is SNAR and what problem does it solve?"@en ; schema:acceptedAnswer :a3 . :a3 a schema:Answer ; schema:text "The Semantic Name Authority Repository at the National Library of Nigeria uses Wikibase to build authority control for Nigerian scholarly and cultural works. It addresses VIAF's thin coverage: only 3 African institutions out of 55 countries participate in VIAF. A Nigerian author moves from being a string to an identified entity in a linked graph."@en . :faq4 a schema:Question ; schema:name "What is 'clouded coloniality'?"@en ; schema:acceptedAnswer :a4 . :a4 a schema:Answer ; schema:text "The phenomenon where heritage data processed in external cloud infrastructure under foreign legal frameworks reproduces the extraction dynamics of colonialism in digital form. The Dar Gnawa Museum's response: build AI tools independently, store data in national data centers."@en . :faq5 a schema:Question ; schema:name "Why don't standard ontologies work for African oral literature?"@en ; schema:acceptedAnswer :a5 . :a5 a schema:Answer ; schema:text "BIBFRAME, MARC, and standard bibliographic ontologies assume knowledge is written, individually authored, and stable. A griot's performance isn't a 'work.' A praise poem that evolves with each telling has no single 'expression.' Collective storytelling resists individual authorship. Forcing these into European categories distorts the thing being described."@en . :faq6 a schema:Question ; schema:name "What is the vindication vs. construction narrative?"@en ; schema:acceptedAnswer :a6 . :a6 a schema:Answer ; schema:text "The European linked data story is retrospective: libraries built infrastructure before anyone knew AI needed it. The African story is simultaneous: authority files and AI applications being built in the same generation, with purpose visible from the start. The builders can see why their work matters as they do it."@en . :faq7 a schema:Question ; schema:name "How does linked data support data sovereignty?"@en ; schema:acceptedAnswer :a7 . :a7 a schema:Answer ; schema:text "Linked data is decentralized by design. A national library can maintain its own authority files in its own infrastructure while federating with global systems through standardized protocols. The data stays where the institution governs it — structurally more compatible with sovereignty than centralized cloud models."@en . :faq8 a schema:Question ; schema:name "What AI tools are African libraries building?"@en ; schema:acceptedAnswer :a8 . :a8 a schema:Answer ; schema:text "AI-assisted subject indexing tools using open-source Python and Hugging Face models trained on African datasets — lighter weight, lower infrastructure requirements. Query response times reduced from days to milliseconds. The cataloging work happening now is the foundation these tools run on."@en . :faq9 a schema:Question ; schema:name "Why does VIAF's African coverage matter for AI?"@en ; schema:acceptedAnswer :a9 . :a9 a schema:Answer ; schema:text "VIAF is the backbone of international bibliographic interoperability with 60M records — but only 3 African institutions participate. When African authors aren't disambiguated in VIAF, AI systems trained on global discovery data don't know they exist. Entity absence in authority files cascades into AI knowledge gaps."@en . :faq10 a schema:Question ; schema:name "What is the Dar Gnawa Museum's Kouyou?"@en ; schema:acceptedAnswer :a10 . :a10 a schema:Answer ; schema:text "An AI tool independently built by the Dar Gnawa Museum in Marrakech, storing all heritage data in certified national data centers in Morocco. A deliberate data sovereignty choice — rejecting the alternative of allowing heritage data to be processed in external cloud infrastructure."@en . :faq11 a schema:Question ; schema:name "How does this change the author's European argument?"@en ; schema:acceptedAnswer :a11 . :a11 a schema:Answer ; schema:text "It transforms it from a vindication narrative about European libraries into a global story. African institutions are building the same linked data infrastructure — but with the purpose visible from the start, with higher stakes (cultural survival, not just discovery efficiency), and with hard-won lessons about data sovereignty built in from day one."@en . :faq12 a schema:Question ; schema:name "What should European GLAM practitioners learn from this?"@en ; schema:acceptedAnswer :a12 . :a12 a schema:Answer ; schema:text "That the 'libraries as AI infrastructure' story is global, not European. That ontology design can be an act of cultural sovereignty. That the question of who controls semantic infrastructure determines whose knowledge counts in AI systems. And that the most interesting chapters are being written in places the Eurocentric professional map barely marked."@en . :glossarySet a schema:DefinedTermSet ; schema:name "Glossary of Key Terms"@en ; schema:hasDefinedTerm :linkedDataInfrastructure, :dataSovereignty, :africanGLAM, :authorityControl . :howtoSection a schema:HowTo ; schema:name "How to Support African GLAM Linked Data Infrastructure"@en ; schema:description "Five actions for global GLAM practitioners, funders, and technologists to support the linked data and AI infrastructure being built by African institutions."@en ; schema:step :step1, :step2, :step3, :step4, :step5 . :step1 a schema:HowToStep ; schema:name "Interrogate Your Metadata Frameworks for Eurocentric Bias"@en ; schema:position 1 ; schema:text "Audit the ontologies, authority files, and cataloging standards your institution uses. Ask: what forms of knowledge are they built to describe? What forms are they built to exclude? MARC, BIBFRAME, and VIAF carry assumptions about authorship, work, and text that systematically exclude oral traditions and collective cultural production."@en . :step2 a schema:HowToStep ; schema:name "Amplify African GLAM Work Without Appropriating It"@en ; schema:position 2 ; schema:text "Share and cite the work of ALMEDA, SNAR, Dar Gnawa, and other African initiatives. Link to their infrastructure from global discovery systems. But do not absorb their data into frameworks that strip its context — federation, not extraction."@en . :step3 a schema:HowToStep ; schema:name "Support Ontology Pluralism"@en ; schema:position 3 ; schema:text "Accept that a single global ontology cannot describe all human knowledge. Support the development of domain-specific and culturally-specific ontologies that coexist through linked data protocols. ALMEDA's decision to build from the epistemological ground up is not a workaround — it is the correct approach."@en . :step4 a schema:HowToStep ; schema:name "Fund African Data Sovereignty Infrastructure"@en ; schema:position 4 ; schema:text "Direct funding toward African institutions building their own infrastructure — not toward solutions that process African data in foreign cloud environments. Support national data centers, local Wikibase instances, and community-governed authority files. Fund the builders, not the extractors."@en . :step5 a schema:HowToStep ; schema:name "Include African Voices in Global Standards Bodies"@en ; schema:position 5 ; schema:text "VIAF, W3C, Dublin Core, and other standards bodies must include African institutions as equal participants — not as recipients of exported standards. The fact that VIAF includes only 3 of 55 African countries is a structural problem that standards bodies must address directly."@en . :kgGeneratorSkill a schema:SoftwareApplication ; schema:name "kg-generator skill"@en ; schema:url ; schema:description "LLM-prompt-based Knowledge Graph generation skill."@en .