National Archives — What's not online
Original records can exist without being available to view online; a search index is not the entire archive.
Why the boundaries of an AI model's training material, searchable records and retrieval behaviour cannot be treated as the boundaries of everything people know.
← Emerging Research Horizon · Interdisciplinary Atlas · Information Science
AI training data and retrieval capabilities do not define the limits of human knowledge. Human knowledge includes embodied expertise, social practice, observation, testimony, unreported findings, protected archives, local and Indigenous knowledge, and work not represented—or not retrievable—within a particular AI system.
A careful argument distinguishes existence from documentation, accessibility, representation and retrievability. Missing search results do not prove nonexistence; nor does a newly encountered claim automatically establish truth or innovation. Claims still require appropriate evidence and testing.
Logical consequence: an AI response reflects a conditional evidence-and-inference process, not a census of all human knowledge. The boundaries are different at every stage.
Archival example: The National Archives confirms that not all records can be viewed online. Therefore an unsuccessful web search cannot establish that no relevant historical record exists.
Scientific example: The WHO describes incomplete clinical-trial results reporting. A literature search can miss existing experimental findings because those findings were not disclosed.
Social example: UNESCO recognises oral and Indigenous knowledge traditions, while OECD documents digital disparities. Communities may possess and apply knowledge that is poorly represented in searchable digital sources.
Technical example: Google Research (2026) separated encoded facts from recall failures in frontier models; the reported percentages concern its evaluated facts and methodology, not a claim that a model has encoded all possible knowledge.
Original records can exist without being available to view online; a search index is not the entire archive.
Recorded life stories, voice collections and restricted-access archives hold testimony not necessarily indexed as searchable text.
Cultural and social knowledge travels through oral traditions and collective memory, not solely published documents.
Long-term observations and locally maintained practices can contribute to discovery and may be historically marginalised.
Local knowledge includes written, oral, visual, tacit, practical and scientific forms, with differing validation traditions.
Scientific practice can depend on socially acquired tacit expertise that is difficult to communicate through explicit texts.
Research impact and practitioner skill often depend on know-how not fully represented by formal publications.
Completed research may be unreported, producing gaps and biases in the accessible evidence base.
Unpublished work and language restrictions affect the completeness and fairness of evidence syntheses.
Access, skills and participation vary across populations; the online record is not equally representative of every group.
Languages are unequally represented in digital technology, shaping whose knowledge is discoverable.
Research should embrace multiple knowledge systems and make knowledge creation accessible beyond traditional scientific communities.
Testimony is a central source of human knowledge and requires critical examination of justification, credibility and context.
Social position and prejudice can affect whose testimony, questions or contributions receive consideration.
A model's inability to recall a fact need not imply the fact was absent from its learned representations.
Retrieval and attribution can improve scientific synthesis but do not eliminate retrieval, coverage or citation failures.
Evaluation datasets may fail to represent full real-world use; relevant stakeholders and contextual testing remain necessary.
Whether knowledge exists is an ontological question; how a person or system knows it is an epistemological question. The fact that systems are incomplete is a reason for inquiry, not a licence to accept extraordinary or unsupported claims. Human testimony, Indigenous knowledge, archival material and expert practice deserve serious assessment, not automatic acceptance or dismissal. Some knowledge must remain private or protected by rights, consent, confidentiality and cultural stewardship.
Information scientists study indexing and retrieval; social scientists study unequal access and credibility; historians and archivists investigate preservation; anthropologists examine situated knowledge; AI engineers assess coverage, recall and error; legal and policy researchers examine privacy, ownership and accountability. Effective governance needs contributions across these perspectives rather than equating a system's available records with a final picture of reality.
Research questions and evidence-gaps register · Methods and mapping · Recognition and evidence
Established 5 November 2026 · 17 independently linked research and institutional resources · Curated interpretation, not systematic literature review or validation of any particular undisclosed discovery.