AI Disclosure Statement Templates for Dissertations and Articles (6 Formats + Publisher Guide)

What Is an AI Disclosure Statement

An AI disclosure statement is a declaration that tells readers, reviewers, and editors exactly how artificial intelligence tools were used in your research paper, thesis, or dissertation. It does not admit wrongdoing. It is simply the new standard for academic transparency.

The statement covers four things: which AI tools you used, what version or date, what specific task each tool performed, and how you verified the output. It ends with a responsibility statement that confirms the final content is your own work.

This requirement has shifted from detection-based compliance to declaration-based compliance. More than 50 universities disabled Turnitin’s AI writing indicator between 2023 and 2026, including Vanderbilt, Waterloo, Curtin, and Washington State. The stated reason was unreliable detection scores, especially for non-native English speakers. Rather than penalising AI use through detection, institutions moved to a disclosure model: the problem is not using AI, it is failing to mark it. Per German legal guidance issued in January 2026 by Baden-Württemberg, the deception is not using AI, it is failing to disclose it.

If you are submitting a thesis, the disclosure goes where your graduate school’s policy says. If you are submitting to a journal, the disclosure goes where the publisher’s policy says. The two are separate requirements. Satisfying one does not satisfy the other.

When Disclosure Is Required

The consensus across major publishers and ethics bodies is clear: AI cannot be listed as an author. ICMJE (International Committee of Medical Journal Editors) and the ACM all agree on this principle. The “disclose when in doubt” rule (Cleland et al., 2026, AMEE Guide No. 192) underpins the modern disclosure framework see source. COPE (Committee on Publication Ethics) reinforces this transparency mandate. Core disclosure elements include the tool name, version or date, specific task, affected section, human verification, and a responsibility statement.

Most institutions distinguish between light editing and substantive assistance. Grammar and spelling checking is generally exempt from disclosure. Substantive editing, translation, data analysis, code generation, and content structuring all require disclosure. Source verification is mandatory for all AI-generated citations regardless of the tool used. If your AI tool generated citations, learn how to cite AI chat conversations properly across APA, MLA, and Chicago styles to avoid formatting errors that undermine your credibility.

For students who are already exploring AI use, see our guide on how to use AI tools ethically in academic writing for the full ethical framework. If your paper has been flagged for language issues by a reviewer, understanding when a language editor is necessary and how to disclose it is equally relevant to the disclosure conversation. ThesisAI’s institutional policy overview provides additional context on the different types of policies universities enforce, including prohibited, permitted-with-disclosure, and permitted-for-defined-tasks categories.

Six Copy-Paste AI Disclosure Templates

Different situations call for different levels of disclosure. Below are six complete, copy-paste-ready templates ranging from “no AI used” to “heavy substantive assistance.” Each includes placeholders for tool name, version, dates, chapter or section numbers, and verification steps. Replace the bracketed sections with your specific details before submission.

Template 1: No AI Used (Standard Declaration)

Use this when you did not use any AI tools during the preparation of your work. Many graduate schools require a declaration even when no AI was involved, simply to confirm that the author has considered the issue.

AI Disclosure Statement

I confirm that no artificial intelligence (AI) tools or AI-generated content were used in the preparation of this [thesis/dissertation/research article]. All text, figures, tables, code, and references were produced solely by the author(s) listed above. Any assistance from human editors or proofreaders was limited to language review and was disclosed separately below.

Some institutions require this declaration even when no AI was used. If your graduate school or journal mandates a mandatory declaration, this template satisfies that requirement with zero assumptions.

Template 2: Grammar and Spelling Only (Exempt Category)

Use this when you used an AI grammar tool such as Grammarly, LanguageTool, or Quill, but no substantive editing or content generation was involved. Most publishers and universities treat basic grammar checking as exempt from formal disclosure, but a short statement is still recommended.

AI Disclosure Statement

Artificial intelligence tools were used for grammar and spelling checking only in this [thesis/dissertation/research article]. The tool [Tool Name, e.g., Grammarly Premium] was used to review grammar, spelling, and punctuation throughout the manuscript. No AI-generated text, content structuring, data analysis, or translation was produced by AI tools. All substantive content, including the literature review, methodology, analysis, discussion, and conclusion, was written solely by the author(s). The author(s) take full responsibility for all text, figures, tables, and references in this work.

This template satisfies institutions that require a declaration even for exempt AI use. It is honest about the tool used without inflating the role of AI in the work.

Template 3: Brainstorming and Outlining (Light Substantive Use)

Use this when you used AI to generate research ideas, suggest structure, or produce an outline, but all substantive writing and analysis are your own.

AI Disclosure Statement

Artificial intelligence tools were used in the preparation of this [thesis/dissertation/research article] as follows: The tool [Tool Name, e.g., ChatGPT-4] was used on [dates] for brainstorming research topics and suggesting structural outlines for Chapters [X–Y]. The tool was used for initial idea generation only; all subsequent writing, analysis, data interpretation, and revision were performed exclusively by the author(s). No AI-generated text was included in the final manuscript without careful verification and rewriting by the author(s). The author(s) confirm that all cited sources were verified as real and accurately represented. The author(s) take full responsibility for all content in this work.

This template covers a common and legitimate use of AI in academic work. The key distinction is that the outline was produced by AI but the writing that followed was human. Reviewers and supervisors should recognise that structural suggestions are not the same as generated content.

Template 4: Coding and Data Analysis Support

Use this when you used AI tools to generate, debug, or verify code, or to assist with data processing and statistical analysis. This template is common among computational disciplines (computer science, engineering, data science, bioinformatics).

AI Disclosure Statement

Artificial intelligence tools were used in the preparation of this [thesis/dissertation/research article] as follows: The tool [Tool Name, e.g., GitHub Copilot / ChatGPT] was used on [dates] to assist with code generation, debugging, and verification for [describe specific code, e.g., the Python data-processing pipeline in Chapter 3]. The tool was used to generate boilerplate code, suggest debugging strategies, and verify syntax. All code was reviewed, tested, and modified by the author(s). All data analysis, statistical interpretation, and results were performed exclusively by the author(s). The author(s) confirm that all cited sources were verified as real and accurately represented. The author(s) take full responsibility for all content in this work.

This template is specific enough to satisfy computational departments that often require code-level disclosure. It does not claim the AI produced the final code; it claims the AI assisted with code generation under human supervision. That distinction matters for reproducibility audits.

Template 5: Substantive Editing and Translation Assistance

Use this when you used AI for paragraph-level editing, restructuring, translation, or significant reformatting of your draft. This is the template most international authors need, and it carries the highest disclosure obligation.

AI Disclosure Statement

Artificial intelligence tools were used in the preparation of this [thesis/dissertation/research article] as follows: The tool [Tool Name, e.g., ChatGPT-4 / Claude] was used on [dates] for substantive editing, including paragraph-level rewriting, sentence restructuring, and language translation from [source language] to English. The tool was used to revise Chapters [X–Y] and to improve readability of the Introduction and Discussion sections. All original ideas, data, analysis, interpretation, and conclusions remain the exclusive work of the author(s). All AI-assisted text was carefully reviewed for factual accuracy, source verification, and alignment with the author's argument. The author(s) confirm that all cited sources were verified as real and accurately represented. The author(s) take full responsibility for all content in this work.

This template is the one most students and researchers overlook. International authors often use AI for language polishing without understanding that most publishers now classify substantive editing as a disclosure category. If your editor used AI to restructure paragraphs, the disclosure must say so.

Template 6: Full Comprehensive Disclosure (Maximum Transparency)

Use this when you want to cover every possible use of AI across the project, or when your institution explicitly requires a full audit of AI involvement. This template is useful for PhD dissertations, multi-chapter theses, or journal manuscripts that will undergo a peer-review process.

AI Disclosure Statement

Artificial intelligence tools were used in the preparation of this [thesis/dissertation/research article] as follows:

1. [Tool Name, e.g., ChatGPT-4] was used on [dates] for [specific task, e.g., brainstorming research questions, suggesting structural outlines, and paragraph-level editing in Chapters 2 and 4].
2. [Tool Name, e.g., GitHub Copilot] was used on [dates] for [specific task, e.g., generating boilerplate code in the data-processing pipeline].
3. [Tool Name, e.g., Grammarly] was used for [specific task, e.g., grammar and spelling checking throughout the manuscript].

All AI-assisted text was reviewed, fact-checked, and rewritten by the author(s). All cited sources were verified as real and accurately represented. No AI-generated content was included in this work without human verification. The author(s) take full responsibility for all content, analysis, and conclusions in this work.

AI Usage Logs are available upon request and stored at [location, e.g., supplementary materials / GitHub repository / institutional repository].

This template provides the most complete disclosure and satisfies institutions that reserve the right to request prompt logs. Several institutions including KU Leuven, the University of Georgia, and Heidelberg now allow supervisors or committees to request full AI usage logs. Including a placeholder for log availability in the statement is a small step that signals good-faith transparency.

Publisher Comparison Table (14 Major Publishers)

If you are converting a thesis chapter into a journal article, the disclosure requirements change. Graduate-school disclosures and journal disclosures are separate, and each publisher has its own placement rules and exemptions. Thesify’s comprehensive 2026 policy matrix tracks authorship policies, disclosure locations, grammar exemptions, image rules, peer-review rules, and legal risk warnings across the academic publishing industry. Use it as a pre-submission checklist.

The 14-Publisher Matrix

The table below covers 14 major publishers as documented by Thesify. Thesify’s matrix is one of the most up-to-date resources for tracking these requirements and is updated regularly as policies evolve.

The key takeaway from this table is that placement varies dramatically across the major publishers. Elsevier requires a dedicated declaration before references, which is the most rigid format among the big five. Nature and Science place disclosure in the Methods section, treating AI assistance as part of the methodological transparency. Wiley and Taylor & Francis often accept it in Acknowledgments for light use. ICMJE requires it in both the cover letter and the submitted manuscript, making it one of the strictest disclosure policies in medical publishing. Always check the publisher’s current policy document before submission, because these rules are frequently updated. Thesify’s 2026 policy matrix is updated regularly and serves as one of the most comprehensive tracking resources available.

Understanding the Pre-Submission Checklist

Beyond the placement location, publishers also require a pre-submission checklist. Thesify’s 2026 matrix includes a detailed compliance checklist that covers authorship verification, AI-generated image audits, co-author AI use tracking, and verification of all AI-assisted content. The image audit protocol is particularly important because AI-generated figures and diagrams require separate disclosure, and some publishers require the author to verify that the image was not generated solely by AI without human oversight.

Placement Conventions: Where Does the Statement Go

There are four main placement conventions across institutions and publishers. The location you choose depends on whether you are submitting a thesis or a journal article, and which publisher or department policy applies.

1. Dedicated AI Declaration Section

Many journals and some graduate schools require a dedicated section titled “Declaration of AI Use” or “AI Disclosure Statement.” This section appears before the References or Acknowledgments. It is the cleanest option because it separates the disclosure from the content. Publishers like Elsevier, Frontiers, and MDPI often require this format.

2. Methods Section

Some journals, including Nature, Science, and many ICMJE-compliant journals, prefer AI disclosure embedded in the Methods section. This placement treats AI assistance as part of the methodological transparency, similar to how software tools, statistical packages, or laboratory equipment are disclosed. This format is common in the sciences and medical journals.

3. Acknowledgments

Light AI use, particularly grammar checking and language polishing, is often disclosed in the Acknowledgments section. This placement signals that the AI role is minor and does not warrant a dedicated section. Wiley, Taylor & Francis, and SAGE often accept this format for exempt categories.

4. Appendix or Supplementary Materials

Some institutions allow detailed AI usage logs to be placed in the appendix or supplementary materials. This format is useful for computational disciplines that require code-level transparency. The AI-Cards framework, developed by the AI-Cards team, promotes an “AI Usage Card” appendix concept that tracks tool names, versions, dates, and tasks for each chapter. ThesisDraft’s LaTeX appendix format is another well-regarded template for technical students who need to include detailed logs alongside their manuscript.

Common Mistakes to Avoid

Even experienced authors make mistakes when writing AI disclosure statements. The following table summarises the most common errors and how to avoid them.

The “disclose when in doubt” rule from the AMEE Guide No. 192 is the single most important principle for AI disclosure compliance. Over-disclosure does not damage a paper. Missed disclosure that surfaces after publication, especially if it involves fabricated citations or unverifiable data, is one of the most common triggers for retraction. The deception is not using AI, it is failing to mark it. MasterMindPhD’s template collection (which draws from University of Manchester and Birmingham policies) includes a similar “over-disclosure is safer than under-disclosure” principle that has been widely adopted by graduate schools across the UK and Germany.

The International-Author Angle: Why Disclosure Matters More for Non-Native Writers

This topic is especially relevant for Essays-Panda’s international audience. A landmark Stanford study by Liang et al. (2023), widely cited by universities including Washington State and Vanderbilt, found that AI detectors produced false-positive rates of 61.22% on authentic writing by non-native English speakers across seven different detectors see study. The same detectors produced false-positive rates of only about 3–5% for native English speakers. This disparity is well documented in detection bias research.

This disparity exists because AI detectors look for patterns associated with machine-generated text, but those same patterns appear in authentic writing by non-native speakers: simplified syntax, reduced idiomatic complexity, and structural choices that reflect second-language patterns. The result is that an international author writing naturally in English may be falsely flagged by detectors. Some authors wonder whether running text through machine translation back-and-forth would hide AI-generated patterns. Evidence suggests it does not reliably mask the structural and semantic markers that detectors look for, but the research focuses on human-written non-native English, not translated AI text specifically.

For these authors, disclosure is safer than hiding. If a paper is flagged by a detector and you have already disclosed your AI use, the disclosure becomes your defence. If you have not disclosed, the lack of declaration looks suspicious regardless of whether the AI flag was accurate.

Turnitin’s AI detector officially supports only seven languages: English, Spanish, Portuguese, French, German, Italian, and Japanese. Accuracy drops significantly for other languages. Multiple universities disabled or restricted Turnitin’s AI detector citing false-positive harms to non-native writers. The detection-to-disclosure paradigm shift happened precisely because institutions recognised that detection alone cannot be trusted across language boundaries.

Practical Steps for Non-Native Authors

For international authors submitting to journals, the disclosure statement should be written in clear, verifiable language that reviewers can assess without relying on detector scores. If you used AI translation tools, document the source language and the verification steps you took. If you used AI grammar tools, name the tool and the sections reviewed. If you are unsure whether a specific tool falls into the exempt or disclosure category, follow the “disclose when in doubt” rule and disclose it. The academic community prefers transparency over perfection.

Final Checklist

Before submitting your paper or thesis, run through this checklist to ensure your AI disclosure is complete and compliant:

• Did you use AI for anything beyond grammar and spelling checking? If yes, disclose it.
• Did you name the specific tool, version or date, task, and section in your statement?
• Did you verify every citation and data point that AI produced or suggested? (Use the four-step link verification guide to check DOIs, Crossref metadata, and PubMed PMIDs.)
• Are your thesis disclosure and journal disclosure written separately (if you are submitting to both)?
• Did you follow the “disclose when in doubt” rule if you are unsure?
• Is your placement correct for the target journal or graduate school?

When in doubt, disclose. The academic community is not trying to punish AI use. It is trying to ensure transparency, accountability, and trust. An honest disclosure strengthens your paper. A missing disclosure weakens it.