
Generative AI tools are now part of everyday academic and professional work. Researchers use them to polish language, summarise papers, write code, brainstorm ideas, and draft text. As these tools spread, journals, universities, and funders increasingly expect people to be open about how they use them. Knowing how to disclose AI use properly protects your credibility and helps readers trust your work.
This guide explains why AI disclosure matters, what usually needs to be disclosed, where and how to write a disclosure statement, examples you can adapt, and common mistakes to avoid.
Table of Contents
Why AI Disclosure Matters
- Transparency: readers deserve to know how a piece of work was produced
- Accountability: authors remain responsible for every word, figure, and citation
- Research integrity: undisclosed AI use may be treated as a breach of policy
- Reproducibility: describing AI use in methods helps others understand and repeat the work
- Compliance: many journals and universities now require disclosure
Major publication ethics bodies agree that AI tools cannot be listed as authors, because they cannot take responsibility for the work. Instead, AI use is treated as something to be disclosed.
What Usually Needs to Be Disclosed?
Policies vary, but the general principle is that substantive use of generative AI should be disclosed, while routine tools often do not need to be.
| Type of Use | Usually Disclose? | Notes |
|---|---|---|
| Spelling and grammar checkers | Often not required | Check your journal’s policy |
| AI language editing or rewriting | Often yes | Especially when text is substantially rewritten |
| Drafting text | Yes | Describe which sections and how output was checked |
| Data analysis or coding help | Yes | Usually in the methods section |
| Literature searching or summarising | Often yes | Verify every source |
| Generating images or figures | Yes, and often restricted | Many journals limit AI-generated images |
Check the Rules First
Before writing a disclosure, read the relevant policies:
- Journal or publisher guidelines: often found in author instructions or editorial policies
- University policies: especially for theses, dissertations, and coursework
- Funder rules: some restrict AI use in grant applications or peer review
- Conference guidelines: many have their own AI policies
When policies differ, follow the strictest one that applies, or ask the editor or supervisor for guidance. Because AI rules are being updated frequently, it is wise to check them again just before submission.
What to Include in an AI Disclosure
A clear disclosure usually answers these questions:
- Which tool? Name the AI system and, where possible, the version
- When? The approximate date or period of use
- For what purpose? Such as language editing, summarising, coding, or idea generation
- Which parts? The sections, analyses, or materials involved
- How was it checked? How authors reviewed, edited, and verified the output
- Responsibility: A statement that the authors take full responsibility for the content
Where to Place Your Disclosure
Methods Section
If AI was used in research processes, such as data analysis, coding, transcription, or literature screening, describe it in the methods so others can understand and reproduce your approach.
Acknowledgements or Dedicated Statement
Many journals ask for a separate “Declaration of generative AI use” or similar statement, often placed before the references or in the acknowledgements, for writing assistance.
Cover Letter
Some journals also ask authors to mention AI use in the cover letter to the editor.
Theses and Assignments
Universities may require a declaration form, a statement in the introduction or appendix, or a log of prompts. Follow your institution’s specific instructions.
Example Disclosure Statements
The following examples can be adapted to your situation and policies.
Language Editing
“During the preparation of this manuscript, the authors used [tool name and version] to improve the readability and grammar of the text. The authors reviewed and edited all suggested changes and take full responsibility for the content of the publication.”
Coding Assistance
“[Tool name] was used to help write and debug parts of the analysis code in [software]. All code was reviewed, tested, and validated by the authors. The final code is available in [repository].”
Literature Screening
“[Tool name] was used to assist in the initial screening of titles and abstracts. All included studies were independently checked by two authors against the original full texts.”
No AI Use
Some journals ask authors to confirm when no generative AI tools were used: “The authors did not use generative AI tools in the preparation of this manuscript.”
Keeping an AI Use Log
A simple log makes disclosure much easier, especially for long projects such as theses or multi-year studies. A useful log might record:
- The date of use
- The tool and version
- The task, such as summarising an article or editing a paragraph
- Key prompts or instructions given
- How the output was checked or changed
Some universities now ask students to attach this kind of record to their submissions. Even when it is not required, a log helps you answer questions confidently if an editor, examiner, or co-author asks how AI was used.
Disclosure in Grant Applications
Funding agencies have started issuing their own guidance. Some allow limited AI assistance with writing but require applicants to take full responsibility for accuracy and originality. Others warn that AI-generated text may raise concerns about plagiarism or confidentiality. Applicants should read each funder’s current rules carefully and disclose any substantial AI support where requested.
Disclosure in Teaching and Coursework
Students should check each course’s rules, as instructors may set different limits. Some assignments may allow AI for brainstorming but not for writing final answers, while others ban it entirely. When AI use is permitted, a short note explaining how it was used is good practice.
AI Use in Peer Review
Reviewers should be especially careful. Uploading confidential manuscripts into public AI tools may breach confidentiality rules. Many publishers and funders restrict or prohibit the use of AI tools by reviewers, or require disclosure. Always check the journal’s reviewer guidelines first.
Common Mistakes to Avoid
- Vague statements: “AI was used” without saying how or where
- Listing AI as an author: not permitted by major publication ethics bodies
- Failing to verify output: AI can produce errors and fabricated references
- Ignoring confidentiality: entering sensitive data or unpublished work into public tools
- Using AI images without permission: many journals restrict AI-generated figures
- Disclosing too late: declare AI use at submission, not after questions are raised
Best Practices for Responsible AI Use
- Keep a simple record of tools, dates, and purposes as you work
- Check every fact, quotation, and citation produced with AI help
- Protect personal, confidential, and unpublished data
- Discuss AI use with co-authors and supervisors
- Be honest and specific in your disclosure
- Stay updated, as policies are changing quickly
Final Thoughts
Learning how to disclose AI use is now an essential skill for researchers, students, and writers. By checking policies, naming the tool, explaining its purpose, describing how outputs were verified, and taking full responsibility for the work, you can use AI tools openly and ethically. Clear disclosure protects your reputation and helps maintain trust in research.





