Project 01 · Research Note 01

When AI Becomes Evidence

A Comparative Audit of Machine-Signal Disclosure in University Academic Integrity Procedures

Navi Musaget · Foundation for Relational Inquiry (FFRI-NM) · Version 1.0 · Published 21 September 2026
The project

What happens when a machine signal enters a human procedure?

FFRI-NM's first completed research project audits public academic-integrity procedures at 24 purposively selected universities in Australia, Canada, the United States and the United Kingdom.

It examines how AI-detector outputs move from technical signals into institutional evidence, what part of the signal becomes visible to the affected student, when disclosure occurs, and how the signal can be challenged.

Scope and limits

A procedural audit, not a verdict on AI detection.

The study separates tool-layer visibility, formal evidence disclosure, human interpretation, signal-level contestability and decision-level review.

It is descriptive rather than normative. It does not estimate prevalence, rank institutions, assess legal compliance, or test detector accuracy.

Sample
24 purposively selected universities
Jurisdictions
Australia · Canada · United States · United Kingdom
Research object
Public academic-integrity procedures
Version
Frozen publication package v1.0
Companion materials include the public corpus, Codebook v0.4, provenance records, audit trail, checksums and the frozen v1.0 publication package.
Public record

Read it, inspect it, trace it.

The research note is published with its companion materials so that the analytical framework and public-source trail can be inspected alongside the paper.

Zenodo research recordResearch note, companion materials and persistent DOI.
Open →
DOI10.5281/zenodo.22879568
Resolve →
FFRI-NM public repositoryPublic institutional record and verification material.
GitHub →
Open inquiry

Found something we missed?

Project 01 is published as a frozen v1.0 record, but the inquiry does not end with publication. We welcome specific challenges that can be checked against public evidence.

You may have found a public procedural source we missed, a classification that should be reconsidered, an alternative interpretation of the evidence, or a way to replicate the audit in another jurisdiction.

Make the challenge inspectable.
Tell us the institution or jurisdiction, the specific claim or classification you are challenging, and the public source or argument that supports your view. A contribution does not alter the frozen v1.0 package; it may inform an erratum, a documented follow-up, or future research.