How to avoid vague disclosure: an optional verification module for the Standard, and a costed audit pilot for journals

A contribution to the drafting of the Global Reporting Standard for AI Disclosure in Research (the "Vancouver Standard"), offered to the core team while the draft is being written. It answers the question "How to avoid vague disclosure?" and Block 5 of the Round 2 Preparatory Reading (v4.5). To the Standard, it proposes an optional, machine-readable verification module: counts of the checks made, each with its denominator (sources cited and checked, factual claims, adversarial readings, defects found and corrected, and the result of any external audit). To journals, outside the Standard, it proposes a costed, pre-registered audit pilot: one volunteer journal or preprint server, for one year, auditing a random fraction of all articles; thirty claim-bearing sentences drawn from each audited text; an exact one-sided 95% upper bound; and due process for the author. The paper applies the method to its own drafts: three random-draw audits of working versions are reported with their counts and bounds, and the verification documents are deposited with it (mostly in French). Files in English, French and Spanish. A short brief is deposited separately: https://doi.org/10.5281/zenodo.24365352 Use of AI: Claude (Anthropic) was used for documentary research, source checking, drafting from the author's notes, and translation; seven AI systems from other developers reviewed an earlier version. No human has yet audited the text or reviewed the translations. This is an independent contribution; it has not been endorsed by the International Science Council or by the drafting team.

Authors

Publication Details

Journal
Zenodo (CERN European Organization for Nuclear Research)
Published
2026-11-19
DOI
https://doi.org/10.5281/zenodo.24365344
Primary Topic
Academic Publishing and Open Access
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

How to avoid vague disclosure: an optional verification module for the Standard, and a costed audit pilot for journals

Marc-Julien Arenas
Zenodo (CERN European Organization for Nuclear Research)
Academic Publishing and Open Access
article

How to avoid vague disclosure: an optional verification module for the Standard, and a costed audit pilot for journals

Marc-Julien Arenas
article en

Abstract

A contribution to the drafting of the Global Reporting Standard for AI Disclosure in Research (the "Vancouver Standard"), offered to the core team while the draft is being written. It answers the question "How to avoid vague disclosure?" and Block 5 of the Round 2 Preparatory Reading (v4.5). To the Standard, it proposes an optional, machine-readable verification module: counts of the checks made, each with its denominator (sources cited and checked, factual claims, adversarial readings, defects found and corrected, and the result of any external audit). To journals, outside the Standard, it proposes a costed, pre-registered audit pilot: one volunteer journal or preprint server, for one year, auditing a random fraction of all articles; thirty claim-bearing sentences drawn from each audited text; an exact one-sided 95% upper bound; and due process for the author. The paper applies the method to its own drafts: three random-draw audits of working versions are reported with their counts and bounds, and the verification documents are deposited with it (mostly in French). Files in English, French and Spanish. A short brief is deposited separately: https://doi.org/10.5281/zenodo.24365352 Use of AI: Claude (Anthropic) was used for documentary research, source checking, drafting from the author's notes, and translation; seven AI systems from other developers reviewed an earlier version. No human has yet audited the text or reviewed the translations. This is an independent contribution; it has not been endorsed by the International Science Council or by the drafting team.

Zenodo (CERN European Organization for Nuclear Research)
Academic Publishing and Open Access
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon