What question did this study set out to answer?

This project aims to explore the causes of AI governance failures related to representation discrepancies.

June 17, 2026Open Access

Representation Integrity: Why Most AI Governance Failures Begin Before the Model Runs

Key Points

This project aims to explore the causes of AI governance failures related to representation discrepancies.
Introduces the Representation Integrity Model (RIM) for analyzing AI governance
Proposes six principles to assess representation integrity
Supports findings with figures and supplementary materials.
Identifies six key principles: Provenance, Contextual Integrity, Contestability, Completeness, Temporal Coherence, and Legitimacy.
Highlights the propagation of failures from machine-legible representations into governance issues.
Demonstrates the necessity of addressing representation integrity before model execution.

Abstract

This project contains the preprint, figures, and supporting materials for the paper "Representation Integrity: Why Most AI Governance Failures Begin Before the Model Runs." The paper introduces the Representation Integrity Model (RIM), a framework for understanding how failures in machine-legible representations propagate into downstream governance failures in artificial intelligence systems. The framework proposes six principles of representation integrity: • Provenance • Contextual Integrity • Contestability • Completeness • Temporal Coherence • Legitimacy The project forms part of a broader research program including The Representation Economy, SENSE–CORE–DRIVER, and Digital Anthropology for Enterprise AI. Related Identifiers ORCID 0009-0002-6207-602X Zenodo DOI 10.5281/zenodo.20687762 Figshare DOI 10.6084/m9.figshare.32668665 OSF DOI 10.17605/OSF.IO/E359Y

Read Full Paperexternally

Mark Helpful

Bookmark

Relay

View Full Paper