SciSpace Academy
Tutorial

AI Workflow From Manuscript Writing to Submission and AI Disclosure

Dr. Joseph George ShenekjiEnglishResponsible AI

TL;DR

1-min read

Many researchers already use AI somewhere in their writing, yet fewer are sure a journal would accept that use or how to disclose it. This tutorial sets out a practical workflow from drafting a manuscript to submitting it. It is designed for early-career researchers, PhD students, postdocs and supervisors who want to use AI without putting a submission at risk, and no technical background is needed.

Key takeaways

  • AI can usefully support some drafting tasks, introduces risk in others and should not be used at all in certain parts of a manuscript.
  • Journals draw a firm line between AI assistance and authorship, and AI tools cannot be listed as authors.
  • ICMJE, COPE, Elsevier, Springer Nature and Wiley share core principles but differ on details such as whether basic editing must be declared and where disclosures belong.
  • A useful disclosure statement names the tool, explains what it was used for and at which stage, and confirms that the authors reviewed the output and accept full responsibility.
  • Adaptable example statements make it easier to write a disclosure that suits your target journal.
  • Common mistakes in AI use and disclosure can lead to desk rejection or post-publication correction, so they are worth checking before you submit.
Apply it

To apply this, note how you use SciSpace while drafting, read your target journal's AI policy, and adapt one of the example statements into your manuscript before submission.

How to do it

  1. Sort drafting tasks by AI risk

    Identify where AI supports your drafting, where it adds risk and where you should avoid it entirely. This keeps AI out of the parts of a manuscript where it does not belong.

  2. Note how you use SciSpace

    Keep notes on how you use SciSpace while drafting, including the task and the stage. These details feed directly into your disclosure.

  3. Read your target journal's AI policy

    ICMJE, COPE, Elsevier, Springer Nature and Wiley share core principles but differ on details such as declaring basic editing. Remember that AI tools cannot be listed as authors.

  4. Adapt an AI disclosure statement

    Adapt an example AI disclosure statement that names the tool, what it was used for and at which stage. Confirm the authors reviewed the output and accept full responsibility.

  5. Check for common mistakes before submitting

    Review your AI use and disclosure against common mistakes before you submit. These errors can lead to desk rejection or post-publication correction.

Watch the tutorial48 min

AI in Manuscript Writing: Journal Policies and Disclosure StatementsPlay video: AI in Manuscript Writing: Journal Policies and Disclosure Statements

Chapters

  1. Welcome and agenda
  2. Can ChatGPT help write a research paper? 5 legitimate AI uses
  3. Can AI switch citation styles and summarize papers?
  4. What you cannot use AI for in research: fabricating or interpreting data
  5. Can AI do the statistical analysis for your research?
  6. Can ChatGPT write your results section or peer review?
  7. Why undisclosed AI-generated text is the biggest red flag
  8. AI as assistant vs AI as author in research papers
  9. Can ChatGPT be listed as an author? ICMJE, Elsevier, Springer Nature and Wiley rules
  10. How to write an AI disclosure statement: 4 parts and sample wording
  11. Do you need to disclose AI used as a research tool or for a disability?
  12. AI disclosure statement examples: language editing, drafting and multiple uses
  13. AI mistakes that get your paper desk-rejected
  14. Springer Nature AI policy: core expectations
  15. Green, low-risk AI uses allowed by Springer Nature
  16. Amber, medium-risk AI uses that need human oversight
  17. Red, prohibited AI uses in research publishing
  18. What does Elsevier require when you use AI in a manuscript?
  19. How to verify AI answers with SciSpace Chat with PDF
  20. Using an AI paraphrase tool to polish academic English

Try this on SciSpace

Frequently asked questions

What should an AI disclosure statement include?

A useful statement names the tool and version, explains what it was used for and at which stage, and confirms that the authors reviewed and edited the output and take full responsibility for the content. Follow the target journal's instructions on wording and placement, as publishers differ on where the statement belongs.

Do journals allow AI-assisted manuscript writing?

Most major publishers allow some AI assistance, such as improving language or readability, provided it is disclosed where required and the authors remain accountable. They do not allow AI to be listed as an author, and they generally prohibit using generative AI to create or alter research data or images. Always check the specific journal's policy.

Why do publishers' AI policies differ?

Publishers and ethics bodies such as ICMJE and COPE share core principles, including human accountability and no AI authorship, but they word their requirements differently and vary on details such as whether basic editing must be declared and where disclosures go. Comparing the policies relevant to your field helps you prepare a submission that meets each one.

Related tutorials