Disclosing AI in Academic Writing

If you’ve used ChatGPT to brainstorm an outline, refined a paragraph, or translated text for a course assignment, the question isn’t whether you need to disclose AI use — it’s how. The answer has become more precise than ever. In 2026, failing to disclose AI use correctly is no longer a minor oversight. It’s a desk-rejection risk that can derail months of research, and the rules vary depending on whether you’re writing a course paper, a conference manuscript, or a thesis.

This guide walks you through the current landscape of AI disclosure policies for students and researchers, provides ready-to-use templates, and shows you exactly where and how to place your disclosure statement so it meets every major journal and university requirement.

AI can never be listed as an author across any major publisher’s policy — this is the single most important rule, and violating it triggers immediate rejection.
Two placement options exist: the Methods section for substantive use (content generation, literature summarization, figure creation) and the Acknowledgments section for writing-adjacent tasks (language polish, formatting). The cover letter alone is never sufficient.
A proper disclosure has four elements: which tool was used (with version), where in the manuscript it was applied, what kind of use occurred, and a statement that you reviewed and take full responsibility for the content.
When in doubt, over-disclose. A small over-disclosure does not harm a paper. A missed disclosure that surfaces after publication can become a retraction risk.

What the Rules Say: Two Universal Principles

Before any template or table, there are two principles that every major journal, university, and ethics body agrees on. I found it useful to keep these front and center when I was navigating disclosure for my own coursework.

Rule 1: AI Cannot Be an Author

Whether you’re writing for Nature, Science, Cambridge University Press, or an undergraduate seminar, the answer is consistent: AI tools do not meet the criteria for authorship. Large language models cannot approve a manuscript, take responsibility for its claims, or sign a copyright agreement. The International Committee of Medical Journal Editors (ICMJE), the Committee on Publication Ethics (COPE), and the Council of Science Editors (CSE) all agree on this. Listing ChatGPT, Claude, or Gemini as a co-author triggers an immediate rejection.

Rule 2: AI Use Must Be Disclosed When It Shapes Content

The threshold varies across publishers and disciplines, but the principle is consistent. You must disclose AI use when the tool produced or shaped content in your paper. If the AI only checked grammar or spelling, most policies consider this standard copy editing — no disclosure required. But if the AI helped generate text, summarize literature, create outlines, or produce figures, the disclosure is mandatory.

When Do You Actually Need to Disclose?

This is where students and researchers get tripped up. Not every AI interaction requires a disclosure, but the line between “use” and “non-use” has blurred as tools have become more capable.

You do NOT need to disclose:

Grammar and spelling checkers — Built-in checkers in Word, Google Docs, or tools like Grammarly fall under standard copy editing.
Reference management software — Zotero, Mendeley, and EndNote use AI for metadata extraction, but this isn’t generative AI use.
Web search queries — Searching Google Scholar or Semantic Scholar doesn’t require disclosure.
AI-assisted coding — GitHub Copilot for code is treated differently from LLMs for text in most policies.

You DO need to disclose:

Any text that ends up in the manuscript — Even a single paragraph drafted or substantially rewritten by AI.
Literature summarization — If you asked an AI to summarize 47 papers and incorporated those summaries, even if you rewrote them.
Outlines or section drafts — If AI helped create your paper’s structure or draft sections.
Figures, alt text, or captions — Any visual element generated with AI assistance.
Brainstorming research questions — If AI contributed ideas that shaped your research design.

The “Disclose When in Doubt” Rule

AMEE Guide No. 192 recommends the “disclose-when-in-doubt” approach as the safest position. A small over-disclosure does not harm a paper. A missed disclosure that surfaces after publication can become a retraction risk.

Where to Place Your AI Disclosure

The placement of your disclosure statement is one of the most common reasons for desk rejection. Let me clarify this with concrete examples from the guidelines I reviewed.

Methods Section

When: The AI was involved in substantive research activities — generating text for the paper itself, creating figures, analyzing data, or summarizing literature.

Examples of journals requiring Methods placement:

Nature Portfolio
Science
PNAS
JMIR
Most clinical and medical journals (via ICMJE alignment)

Why it matters: Placing the disclosure in Methods signals that the AI use is part of the research process and is therefore subject to the same scrutiny as any other methodological choice.

Acknowledgments Section

When: The AI was involved only in writing-adjacent tasks — language polishing, formatting, generating title suggestions, or improving readability.

Examples of journals accepting Acknowledgments placement:

Cell Press journals (for language-only use)
Elsevier journals (as their standard requirement)
Springer Nature journals (for language-only use)

Cover Letter Is NOT Enough

This is critical. Several journals explicitly reject papers that disclose AI use only in the cover letter. The cover letter is read by the editor; reviewers and readers do not see it. Disclosure has to live in the published version of the paper.

The Default Rule

When in doubt, place your disclosure in Methods. A Methods disclosure satisfies almost every current policy. An Acknowledgments-only disclosure does not.

The Four Elements of a Proper Disclosure Statement

A complete, policy-compliant disclosure has four elements. I found this framework from the InstaText research template to be the most reliable:

Which tool (with vendor and version)
Where in the manuscript it was applied
What kind of use occurred
A responsibility statement confirming you reviewed and take full responsibility

Here are concrete templates for each use case.

Template 1: Language Editing and Clarity

“During the preparation of this manuscript, the author(s) used [Tool Name, e.g., ChatGPT (OpenAI, GPT-4)] for language editing, specifically to improve clarity and readability in the Discussion section. After using this tool, the author(s) reviewed and edited the content and take full responsibility for the content of the publication.”

Template 2: Content Generation and Summarization

“During the preparation of this manuscript, the author(s) used [Tool Name, e.g., Claude (Anthropic, Claude 3.5 Sonnet)] to summarize 47 candidate papers during the literature review. All summaries were verified against the source papers by the first author before any content was incorporated into the manuscript.”

Template 3: Multiple Tools

“During the preparation of this manuscript, the author(s) used DeepL (version available) for translation from [language] to English and ChatGPT (OpenAI, GPT-4) for language editing and clarity improvements. The author(s) reviewed and revised all outputs from these tools and take full responsibility for the final content of the work.”

Template 4: Coursework / University Assignments

“I acknowledge the use of ChatGPT (OpenAI, https://chat.openai.com/) to refine the academic language and structure of my essay draft. The prompts entered and the resulting outputs were used exclusively to improve the clarity of my original ideas.”

Template 5: No AI Used (When Required)

Some journals require a declaration regardless of whether you used AI. If this applies to your submission:

“I have not used any AI or AI-assisted technologies to prepare this work.”

Major Journal AI Policies in 2026

The landscape has shifted dramatically since 2023. Here’s a practical summary of what every major publisher requires right now. I pulled this from the policies I reviewed at Princeton University, Alfred Scholar, and the publisher documentation itself.

Publisher	Placement	Content Use	Language Use	Notes
Elsevier	Acknowledgments	Required	Not required	Must be inserted above references
Springer Nature	Methods (substantive) / Acknowledgments (language)	Required	Not required	Follows Nature policy
Nature Portfolio	Methods	Required	Not required	AI-generated images banned outright
Science	Methods	Required	Not required	Prompts available on request
Cell Press	Methods (content) / Acknowledgments (language)	Required	Not required	Depends on use type
PNAS	Methods (or Acknowledgments if no Methods)	Required	Not required	Standard practice
Taylor & Francis	Methods or Acknowledgments	Required	Not required	Check specific journal
Wiley	Methods	Required	Not required	Updated 2026 policy

The pattern: Substantive AI use goes in Methods. Language-only use goes in Acknowledgments. Pure grammar checking generally requires no disclosure. Always check the specific target journal’s author guidelines.

University and Student Policies

If you’re writing a course assignment or undergraduate thesis, your university’s policy is your primary guide. Here are examples from institutions I reviewed:

Princeton University: Students must confirm AI is permitted by an instructor and disclose the use of AI in any academic work. The Scholarly Integrity webpage maintains updated guidance on “Disclosing Generative AI.”
Monash University: Mandates that students document their usage, clearly delineate human vs. AI portions of an assignment, and may ask for prompts or chat histories to be saved or appended.
The University of Melbourne: Requires acknowledgment where GenAI has been used as a functional tool to assist in creating academic content. The Academic Skills Kit provides templates and guidance.
UCL (University College London): Advises that the use of GenAI must be acknowledged where it has been used as a functional tool to assist in the process of creating academic content, and cited appropriately in the reference list if specific outputs are quoted.

The bottom line: Always check your course syllabus or department’s AI policy before submitting. Your instructor’s rules override any journal policy you might encounter later.

When to Choose a Course Paper vs. a Journal Submission

The distinction matters because the disclosure requirements differ significantly.

Factor	Course Assignment	Journal Submission
Primary guide	Instructor’s syllabus / department policy	Target journal’s author guidelines
Placement	Usually at end, beginning, or assignment cover sheet	Methods, Acknowledgments, or declaration block
Template needed	Informal acknowledgment acceptable	Formal, structured disclosure required
Transcripts	Sometimes requested by instructor	Often required by medical journals (ICMJE)
Consequence of error	Grade impact, possible academic integrity review	Desk rejection, retraction risk, career impact

Keeping an AI Use Log

Several journals — particularly JMIR, BMJ, and some Elsevier titles — now ask authors to retain full prompts and AI outputs on file in case reviewers want to see them. For students, this is equally practical.

What to log:

The name and version of every AI tool used
The date of each session
The specific section of the paper it was applied to
One-line summary of the use

A practical approach I recommend: Keep a single ai_log.md file in your project folder noting each tool, date, section, and a one-line summary of the use. This is the same hygiene you already apply to lab notebooks and analysis code.

A Pre-Submission AI Disclosure Checklist

Before you submit anything — whether it’s a course paper or a journal manuscript — run through this checklist:

Identify every AI tool used in the project, including writing assistants, literature search tools, and figure generators.
Decide which uses require disclosure under your target’s policy. Use the “disclose-when-in-doubt” rule for borderline cases.
Write the disclosure paragraph and place it in Methods (default) or Acknowledgments (language-only use).
If journal policy requires, mention AI use in your cover letter as well.
Save full transcripts for any AI session that produced manuscript content.
Confirm no AI tool is listed in the author list.
Double-check the word count — declarations are usually not counted in the manuscript word limit, but verify the specific journal’s or instructor’s rules.

What We Recommend: Common Mistakes to Avoid

Based on our review of AI disclosure guidelines from major publishers and university policies, here are the most common pitfalls I saw students and researchers make:

Disclosing only in the cover letter — Reviewers and readers don’t see the cover letter. The disclosure must be in the manuscript.
Listing AI as an author — This is an automatic rejection across every publisher.
Using vague language — “AI was used in the preparation of this manuscript” is insufficient. You must specify which tool, what it did, and confirm your oversight.
Ignoring journal-specific requirements — A disclosure that satisfies Elsevier may not satisfy Nature. Always check the target journal’s guidelines.
Forgetting to check course syllabi — Your instructor’s policy is the first standard you must meet.
Not keeping records — Some journals require full prompt logs. Keeping an AI use log protects you if your work is ever scrutinized.

Final Checklist Before Submission

[ ] Have I identified every AI tool used in the project?
[ ] Is my disclosure placed in the correct section (Methods or Acknowledgments)?
[ ] Does my disclosure include all four elements (tool name, version, specific use, responsibility)?
[ ] Have I verified the target journal’s or instructor’s specific requirements?
[ ] Have I saved full transcripts for any AI session that produced manuscript content?
[ ] Have I confirmed no AI tool is listed as an author?

Next Steps

AI disclosure is one of those skills that separates careful scholars from the rest. The rules are evolving quickly — new frameworks like GAIDeT and the AID (AI Disclosure) Statement Builder are being rolled out by institutions and publishers. Building the habit of full disclosure and complete record-keeping now makes future policy shifts painless.

If you’re navigating a complex submission with multiple AI tools involved, or your department has specific disclosure requirements you’re unsure about, our team of qualified writers can help. We review AI disclosure statements, ensure compliance with journal policies, and make sure your transparency is properly documented.

This guide synthesizes best practices from publisher AI policies (Elsevier, Springer Nature, Nature Portfolio, Science, Cell Press, Taylor & Francis), university guidelines (Princeton University, Monash University, University of Melbourne, UCL), ICMJE and COPE cross-publisher guidance, and practical frameworks documented by InstaText, Alfred Scholar, and the GAIDeT/AID disclosure frameworks. All policies are current as of 2026.

Related Guides

FAQ

How do I acknowledge ChatGPT in a course paper?

Check your course syllabus first. Most instructors ask for a brief acknowledgment at the end of the paper or on the assignment cover sheet. A template like “I used ChatGPT to refine the language and structure of my essay draft. All ideas and arguments are my own” is usually sufficient for coursework.

Where does the AI disclosure go in a journal manuscript?

It goes in the Methods section if the AI was used for substantive research activities (content generation, literature summarization, figure creation). It goes in the Acknowledgments section if the AI was used only for language editing. Some journals require a separate “Declaration” heading above the reference list.

Can I use an AI tool and not disclose it?

If the tool was only used for grammar checking or reference management, most policies consider this standard copy editing — no disclosure required. But if the AI helped generate text, summarize literature, create outlines, or produce figures, disclosure is mandatory across all major publishers.

What happens if I forget to disclose AI use?

The consequences depend on where you are in the publication process. If caught during peer review, it may result in a request to revise and resubmit or a desk rejection. If discovered after publication, it can lead to a retraction and damage your academic reputation.

Should I list my AI prompts in the paper?

Most policies don’t require the prompts themselves in the manuscript. However, several journals (JMIR, BMJ, some Elsevier titles) ask authors to retain full prompts on file. Some journals may request the full transcript as supplementary material during peer review.