AI-R07 — AI for Academic Publishing and Editorial Work

Course Code: AI-R07  |  School: School of Artificial Intelligence  |  Cluster: Level 2 — Research & Academic Practice

Level: Intermediate  |  Duration: 6 weeks · 20–26 learning hours  |  Language: English  |  Certificate: Professional Certificate (non-degree)  |  Format: Self-paced with AI support under human supervision

Overview

Scholarly publishing has had to take a position on generative tools quickly, and the position that has emerged across the major editorial bodies is consistent on one point: a machine cannot be an author, because authorship entails accountability that software cannot bear. Almost everything else remains contested, which makes this a course about judgement under unsettled norms rather than about compliance with settled ones.

The course serves two audiences at once, authors and editors, because the obligations are complementary. Authors learn what may legitimately be assisted in manuscript preparation, what must be disclosed and where, and how to respond when a reviewer alleges undisclosed machine use. Editors learn to write a workable journal policy, to handle allegations proportionately, to recognise the signatures of fabricated citations and paper-mill submissions, and to run peer review in an environment where reviewer reports themselves may be machine-generated.

Assessed work is practical throughout: a disclosure statement, a journal policy, an editorial decision on a contested case, and a peer review report written to a standard the learner would be willing to sign.

Learning outcomes

On completion, a successful learner will be able to:

1. State what current editorial guidance permits, requires and prohibits regarding generative tools, and cite the source of each position.
2. Prepare a manuscript with assisted drafting and language editing that meets disclosure requirements.
3. Write a disclosure statement that is specific enough to be useful to an editor and a reader.
4. Draft a journal or press policy on generative tools covering authorship, disclosure, peer review and enforcement.
5. Detect the characteristic signatures of fabricated references, spurious citation and paper-mill submissions.
6. Handle an allegation of undisclosed machine use proportionately, protecting both integrity and the accused author.
7. Write a peer review report that adds evaluative judgement rather than summary.

Who this course is for

Authors preparing work for publication, journal editors and editorial board members, university press staff, research integrity officers, and reviewers.

Prerequisites

Experience of academic writing and publication, or an editorial role. AI-R02 recommended for the citation verification discipline assumed here.

Syllabus

Module 1 — Why a machine cannot be an author

Focus. Authorship as accountability. The convergence of editorial bodies on excluding machine authorship, the residual disagreements about disclosure, and the reasoning behind both.

Lessons. 1.1 Authorship criteria and accountability. 1.2 The convergence on excluding machine authorship. 1.3 Contested territory: language editing, drafting, analysis. 1.4 Reading guidance critically.

Core reading. International Committee of Medical Journal Editors, Recommendations for the Conduct, Reporting, Editing, and Publication of Scholarly Work in Medical Journals (current edition), section on authorship. Committee on Publication Ethics, Authorship and AI Tools position statement (COPE, 2023). H. Holden Thorp, “ChatGPT is Fun, But Not an Author”, Science 379, no. 6630 (2023): 313.

Deliverable. Comparative note on three editorial bodies’ positions, with the point of disagreement identified.

Module 2 — Assisted preparation and legitimate use

Focus. Language editing for non-native writers, structural revision, and summarisation of one’s own prior work, distinguished from generating substantive content or analysis the author cannot defend.

Lessons. 2.1 Language editing and equity considerations. 2.2 Structural revision. 2.3 Generating content you cannot defend. 2.4 The line, and how to tell you have crossed it.

Core reading. Chris Stokel-Walker & Richard Van Noorden, “What ChatGPT and Generative AI Mean for Science”, Nature 614 (2023): 214–216.

Deliverable. Annotated manuscript section showing assisted and unassisted passages with the basis for each classification.

Module 3 — Disclosure that is actually useful

Focus. Writing disclosure that tells an editor what was done, to which part, and how it was checked. The failure of generic statements, and where disclosure belongs in the manuscript.

Lessons. 3.1 What an editor needs to know. 3.2 Placement and wording. 3.3 Why generic statements fail. 3.4 Disclosure in supplementary and data materials.

Core reading. Committee on Publication Ethics, Authorship and AI Tools (2023). World Association of Medical Editors, Recommendations on Chatbots and Generative Artificial Intelligence in Relation to Scholarly Publications (WAME, 2023).

Deliverable. Disclosure statement for a real manuscript of your own.

Module 4 — Detecting fabrication and paper-mill signatures

Focus. Fabricated and misattributed references, citation of retracted work, tortured phrasing, statistical implausibility, and the limits of detection tools. Proportionate response to suspicion.

Lessons. 4.1 Reference fabrication and how to check. 4.2 Retraction and citation hygiene. 4.3 Paper-mill signatures. 4.4 Detection tools and their false positives.

Core reading. Matthew J. Page et al., “The PRISMA 2020 Statement”, BMJ 372 (2021), for the reporting discipline that makes fabrication visible. Ziwei Ji et al., “Survey of Hallucination in Natural Language Generation”, ACM Computing Surveys 55, no. 12 (2023).

Deliverable. Screening report on a supplied manuscript, listing findings and the confidence attached to each.

Module 5 — Editorial policy and handling allegations

Focus. Drafting policy that an editorial office can apply. Handling an allegation of undisclosed use without presuming guilt, given that detection tools are unreliable and disproportionately flag non-native writing.

Lessons. 5.1 Policy structure and scope. 5.2 Peer review: may reviewers use these tools? 5.3 Allegation procedure and burden of proof. 5.4 Correction, expression of concern, retraction.

Core reading. Committee on Publication Ethics, Guidelines on Retraction and Core Practices (COPE, current editions).

Deliverable. Journal policy draft plus a written editorial decision on a contested case.

Module 6 — Reviewing well in the new environment

Focus. What a review should contain, why summary is not evaluation, confidentiality obligations that preclude uploading manuscripts to external services, and writing a report the reviewer would sign.

Lessons. 6.1 Evaluation versus summary. 6.2 Confidentiality and third-party processing. 6.3 Constructive severity. 6.4 Signing your report.

Core reading. Irene Hames, Peer Review and Manuscript Management in Scientific Journals: Guidelines for Good Practice (Oxford: Wiley-Blackwell, 2007), chapters 3–5.

Deliverable. Final submission: full peer review report on a supplied manuscript, with a confidentiality statement.

Assessment

Pass mark 70 per cent. All assessed components must be attempted. Every mark in this course is issued by a human assessor; no assessment outcome is generated automatically.

Rubric criteria

Each assessed artefact is marked against four criteria at four levels (distinction, pass with merit, pass, fail).

1. Fidelity to source: are claims about editorial requirements attributed to the actual guidance rather than to general impression?
2. Specificity of disclosure and policy: could an editorial office apply these documents without further interpretation?
3. Proportionality: does the handling of suspicion protect integrity without presuming guilt or penalising non-native writing?
4. Evaluative quality: does the review make a judgement and support it, rather than summarising the manuscript?

Reading list

Editorial guidance. International Committee of Medical Journal Editors, Recommendations for the Conduct, Reporting, Editing, and Publication of Scholarly Work in Medical Journals (current edition). Committee on Publication Ethics, Core Practices, Guidelines on Retraction, and Authorship and AI Tools (COPE). World Association of Medical Editors, Recommendations on Chatbots and Generative Artificial Intelligence (WAME, 2023).

Commentary. H. Holden Thorp, “ChatGPT is Fun, But Not an Author”, Science 379, no. 6630 (2023). Chris Stokel-Walker & Richard Van Noorden, “What ChatGPT and Generative AI Mean for Science”, Nature 614 (2023).

Practice. Irene Hames, Peer Review and Manuscript Management in Scientific Journals (Oxford: Wiley-Blackwell, 2007). Matthew J. Page et al., “The PRISMA 2020 Statement”, BMJ 372 (2021).

All items are published works identifiable by author, title and publisher. Learners obtain them through an institutional library or the publisher. The Academy does not distribute copyrighted texts.

Academic integrity and use of AI

Generative tools may be used in producing assessed work under three conditions. Use must be disclosed in a short statement appended to each submission, naming the tool and the task it performed. Any factual or technical claim originating from a generative tool must be verified against a citable source before it enters assessed work, and the verification must be evidenced. The analytical judgement in each artefact must be the learner’s own and must be defensible in a short follow-up. Manuscripts supplied for the screening and review exercises are confidential and must not be uploaded to any external service.

Instructor: pending owner confirmation. Pricing: pending owner approval. Reference list verified against publisher records; any later addition is marked for verification before publication.