Australian Journal of Artificial Intelligence and Decision Sciences
Author AI policy · Editor and reviewer AI policy
Author instructions · Reviewer guidance · Manuscript Pre-check
Practical guidance for all six journals
Explain the assistance used. Verify the evidence. Keep scientific responsibility with people.
Guidance 1.0 · Published 22 November 2026 · Supports adopted policy framework 1.1
This guide explains the existing AI and confidentiality framework and the journal policies linked below. Human authors remain accountable for their work; editors and reviewers retain responsibility for scientific assessment.
Identify relevant AI assistance in writing, substantive editing, translation, literature synthesis, coding, analysis and image creation. Name the tool and provider, give the model or version where available or the access date when a fixed version is unavailable, state the purpose and affected material, and explain the human checks actually performed.
A disclosure explains what happened. It does not make fabricated observations, invented references or unsupported analyses acceptable. Review the tool’s terms, rights and confidentiality arrangements before providing research material.
Use this worksheet to describe what actually happened. Material assistance includes AI use in research, analysis, writing, translation, coding or image preparation. Ordinary non-generative spelling tools need not be itemised under the adopted policy. This exception does not cover generative rewriting.
Preparation only. This does not submit a manuscript. Your entries stay in this page on your device. They are not uploaded or saved by this worksheet. Download or copy your declaration before leaving the page, and do not enter participant identifiers or confidential research data.
Check the wording against the work and the journal instructions before adding it to your manuscript.
| Material | What authors should provide |
|---|---|
| Explanatory illustration | Label it as illustrative. Verify the concepts and labels, disclose material AI assistance in the caption or declaration, and check provenance and reuse rights. |
| Data-derived plot or table | Use the actual data and reported methods. Check values, axes, labels, uncertainty and statistical annotations. Keep the data and code or another justified assessment route. |
| Observed research image | Retain originals and document consequential processing. Do not add, remove or generate features to misrepresent observations or create an apparently authentic participant image. |
| AI imaging or synthetic-data research | Describe the method and validation, label generated or reconstructed material accurately, and retain the relevant source inputs. A legitimate research method must not be presented as an unaltered observation. |
Restricted data can be legitimate. Explain the reason and provide a safe supporting example, controlled access or another credible route where appropriate. Do not expose participant information, credentials or sensitive systems merely to satisfy a checklist.
A chatbot response must not replace an original source as factual authority. If AI output is the object of the research, document it as research material and describe how it was obtained and assessed.
Editors and reviewers must perform scientific assessment themselves. Generative AI must not produce their scientific evaluation or determine acceptance, rejection or reviewer recommendations.
Keep records of the tool, permitted task, version, information exposed, relevant terms, responsible human and disclosure. Scientific decisions remain independent of commercial services and fees.
The Manuscript Pre-check works on the author’s device. It checks detectable sections and declarations, numeric citation structure, a small phrasing list, repetition, figure/table references and optional comparison text. It does not determine whether AI was used, detect fabricated research or search a global plagiarism database.
The separate optional DOI lookup sends extracted DOIs to our status service and Crossref, not the manuscript or full reference text. It checks up to 100 unique DOIs and identifies unchecked sources. Available results may be cached for one hour. Missing notices, service failures and sources without DOIs require manual verification.
Preparation reports can help identify items to examine. They do not replace the journal’s independent review or certify that references are reliable, unretracted or scientifically relevant. Automated flags require human assessment.
| Stage | Record and assess |
|---|---|
| Submission package | Manuscript and file versions; authorship, conflicts, ethics, data/code and AI declarations; correctable omissions. |
| References and evidence | Publication records and notices; sources supporting central claims; consistency of data, analyses, tables and figures; access limits and targeted requests for originals. |
| Independent review | Reviewer expertise, identity and conflicts; the materials examined; confidentiality and AI conditions; at least two independent external reports for scholarly articles under the adopted model. |
| Integrity enquiry | The specific concern and supporting evidence; an unconflicted handler; author questions and responses; independent advice or justified institutional referral. |
| Decision and production | Reasons and resolution of material concerns; accepted-version declarations; author proof approval, publication dates, rights and linked scholarly-record updates. |
A phrase match, detector score, writing style or missing metadata is not proof of misconduct. Examine the specific evidence and give affected parties a meaningful opportunity to respond. Distinguish honest error, a disclosure omission, unresolved uncertainty and an established breach.
People whose own conduct or interests are implicated must not control the investigation. Where material concerns remain unresolved, the affected decision or publication may be paused. Corrections, expressions of concern and retractions follow the published scholarly-record policies.
To raise a concern, use the relevant journal’s published contact route and identify the work, issue and supporting evidence. Keep confidential records and participant information out of public comments. Concerns involving the usual recipient require independent handling.
These blank specifications help prepare a secure editorial workflow. Complete case records only in the journal’s approved, access-controlled system. These downloads do not activate a manuscript-submission service.
Read the journal’s own policies and scientific guidance alongside this guide. The adopted framework covers human responsibility, confidentiality, restricted materials, illustrations, investigations and rights under the applicable article licence.
Author AI policy · Editor and reviewer AI policy
Author instructions · Reviewer guidance · Manuscript Pre-check
Author AI policy · Editor and reviewer AI policy
Author instructions · Reviewer guidance · Manuscript Pre-check
Author AI policy · Editor and reviewer AI policy
Author instructions · Reviewer guidance · Manuscript Pre-check
Author AI policy · Editor and reviewer AI policy
Author instructions · Reviewer guidance · Manuscript Pre-check
Author AI policy · Editor and reviewer AI policy
Author instructions · Reviewer guidance · Manuscript Pre-check
Author AI policy · Editor and reviewer AI policy
Author instructions · Reviewer guidance · Manuscript Pre-check
This guide is informed by the official sources below, reviewed on 21 November 2026. They serve different purposes; a policy statement does not itself establish indexing, membership or external approval.
Published article reuse follows the article’s actual licence and separately identified third-party rights. This guide does not narrow CC BY 4.0 permissions or authorise disclosure of confidential, unpublished material.
Guidance version record · Publishing standards · Records and privacy