Framework paper · Series Paper 5

Evidence-Governed Intelligence: Applying Geometry Intelligence to Institutional AI Assurance.

Final Manuscript 1.0. Authored by Tim Jacobs and published by KTS Global through Zenodo on 2026-08-16 under DOI 10.5281/ZENODO.21962425. Represented in Wikidata by Q141100663.

Publication record

Where to read it.

Series relations

How this paper relates to other works in the series.

The Geometry Intelligence Foundational Series is a numbered sequence of framework papers published by KTS Global on Zenodo. The relations below use a fixed relation vocabulary (follows, references) and are also expressed in the machine-readable JSON-LD embedded on this page.

About this paper

Short description.

The paper applies Geometry Intelligence to institutional AI assurance. It defines an evidence-governed intelligence pattern in which every decision-relevant machine output is bound to a structured evidence record whose identity, admissibility, context, authority and lineage can be inspected by an external reviewer. The paper composes the formal model (Paper 2), the black-box conformance framework (Paper 3) and the sovereign federation architecture (Paper 4) onto the assurance obligations typical of regulated institutional settings.

The paper separates AI capability claims from AI assurance claims. Capability claims describe what a system can produce; assurance claims describe how a reviewer can determine, from evidence surfaces alone, whether a produced output is fit for a declared institutional use. The paper states the minimum evidence-record shape and the minimum reviewer-facing surface required to make an assurance claim intelligible, comparable and auditable across institutions.

The paper is authored by Tim Jacobs and published by KTS Global as a report, in English, under a Creative Commons Attribution 4.0 International licence. It is Series Paper 5 of the Geometry Intelligence Foundational Series and immediately follows Sovereign Web4 Federation (DOI 10.5281/ZENODO.21949720). The authoritative source of the paper's content is the Zenodo record linked above; this page is a discovery and citation surface for that record, not a re-hosting of it.

Citation

How to cite this paper.

Plain text

Jacobs, T. (2026). Evidence-Governed Intelligence: Applying Geometry Intelligence to Institutional AI Assurance. Final Manuscript 1.0. KTS Global. https://doi.org/10.5281/zenodo.21962425

BibTeX

@techreport{jacobs2026evidencegovernedintelligence,
  author       = {Jacobs, Tim},
  title        = {Evidence-Governed Intelligence: Applying Geometry Intelligence to Institutional AI Assurance},
  version      = {Final Manuscript 1.0},
  institution  = {KTS Global},
  year         = {2026},
  month        = {August},
  day          = {16},
  doi          = {10.5281/ZENODO.21962425},
  url          = {https://zenodo.org/records/21962425},
  note         = {ORCID: 0009-0008-7130-1448. Wikidata: Q141100663. License: CC BY 4.0. Series Paper 5. Follows 10.5281/ZENODO.21949720. References 10.5281/ZENODO.21907367, 10.5281/ZENODO.21921341, 10.5281/ZENODO.21943877.}
}
Status and boundaries

What this record does — and does not — establish.

Framework-paper status. This paper is the Final Manuscript 1.0 of Evidence-Governed Intelligence, published on 2026-08-16. It is Series Paper 5 within the Geometry Intelligence Foundational Series and follows Sovereign Web4 Federation (DOI 10.5281/ZENODO.21949720, Wikidata Q141086630). Later revisions, if any, will each carry their own DOI and their own Wikidata publication item.

Boundaries. Zenodo publication and DOI registration do not constitute peer review, scientific endorsement, or independent technical verification of the paper's contents. Zenodo is a general-purpose open-access repository; DataCite is a DOI registration authority; neither performs peer review of deposited content. The Wikidata item reflects that a publication event exists; it does not endorse any technical claim inside the paper. The ORCID association confirms that the author's public ORCID record lists this work under 0009-0008-7130-1448; it is not an ORCID technical verification of the paper's contents. Publication on Zenodo does not establish world-first status, category ownership, or IETF endorsement. Institutional AI assurance is a governance posture; this paper does not certify any specific system.
Wikidata publisher delta. Wikidata Q141100663 currently records the publisher of this work as Zenodo (Q22661177). The authoritative Zenodo record and the DataCite DOI registration identify KTS Global (Q138189229) as publisher. This page follows the authoritative Zenodo and DataCite publisher. The Wikidata delta is tracked as a separate out-of-band correction and is stated here only as a boundary note; it is not asserted by this record.

Anyone can independently retrieve the paper, the DOI registration, the Wikidata publication item and the ORCID public record from those systems directly, and compare against the SHA-256 anchors published on the NODE-6 evidence record.