Skip to main content

Pubrica

More Freedom. More Research. More Savings!
Flat 25% Off on expert services, plus additional savings for a limited time.
More Freedom. More Research. More Savings!
Flat 25% Off on expert services, plus additional
savings for a limited time.
Facebook Youtube Linkedin Instagram X-twitter
Get Expert Help
Pubrica

Human Editing vs AI Writing Tools: What You Must Disclose and What Journals Still Reject

High-Impact Journals

Interesting topics

Human Editing vs AI Writing Tools: What You Must Disclose and What Journals Still Reject

Artificial intelligence (AI) has quickly become an integral part of academic writing, helping researchers with language enhancement, literature organization, drafting, summarizing, translating, and citation management. Nevertheless, the increasing application of AI generators has raised issues related to authorship, transparency, originality, confidentiality, and research integrity. [1]

The most important point to be noted here is that the issue is not anymore of human editing against AI writing. Journals start looking into the usage of AI, human responsibility, disclosure of AI when necessary, and verification of the output produced by AI. The current guidelines provided by ICMJE, Elsevier, and Springer Nature focus on human responsibility and transparency of AI. Researchers using AI writing tools for research papers should therefore understand the relevant journal AI policy and disclosure requirements.

ICMJE: AI cannot be an author; humans are still responsible for accuracy and originality; the use of AI must be disclosed; confidentiality should be maintained during peer review.

Elsevier: Permits responsible AI assistance with human supervision; mandates disclosure; forbids AI authorship; reviewers/editors are forbidden from uploading manuscripts into AI software systems.

Springer Nature: Authors remain responsible; AI cannot be an author; AI usage is generally required to be disclosed; reviewers have to maintain confidentiality of manuscripts; AI-assisted copy editing and generative AI use are distinguished.

1. Human Editing vs AI-Assisted Writing

Human editing consists of making improvements to the manuscript without changing the underlying research idea, interpretation, argument, and conclusion of the author. This may consist of grammatical corrections and improvement in sentence structure, style, and clarity. [2]

On the other hand, AI-aided writing may be a simple linguistic correction or generating large chunks of text. It is important for the ethics of publication. The distinction between

human editing vs AI writing tools is increasingly important when determining appropriate AI writing disclosure.

Type of Use

Typical Example

Disclosure Consideration

Basic editing

Grammar, spelling, punctuation

Often does not require disclosure

Language refinement

Improving readability and wording

Depends on journal policy

Structural assistance

Reorganising sections or arguments

May require disclosure

Generative writing

Producing paragraphs or manuscript sections

Generally, requires disclosure

Research assistance

Literature synthesis, data analysis or coding

Usually requires appropriate disclosure

AI-generated figures

Creating or substantially altering research images

Often restricted or subject to specific rules

Elsevier currently states that basic spelling, grammar, and punctuation checks do not require an AI declaration, whereas substantive AI changes to sentence structure or organisation should be disclosed.

2. What Researchers Must Disclose

Disclosure is now a key aspect of journal submissions. According to ICMJE guidelines, the authors should disclose any AI technologies utilised in manuscript preparation and describe how it was utilized. They also state that it is inappropriate to list AI technologies as the authors of any work since AI cannot take responsibility for accuracy, integrity, and originality. [3]

Researchers are advised to verify the target journal’s guidelines prior to submitting a paper and provide information about:

  • ICMJE: Disclosure should include whether AI was used and how it was used, e.g., a simple use of AI to check grammar would be different from the use of AI to create the actual content.
  • Elsevier: Substantive use of AI requires disclosure, but simple spelling, grammar, and punctuation checks will not require any disclosure.
  • Springer Nature: Transparent disclosure of AI-generated content is required, whereas use of AI for copyediting, grammar, and readability will not require any disclosure.

For instance, Elsevier demands that the researcher provide a special declaration of the usage of any generative AI in manuscript preparation. This AI disclosure statement is an important part of transparent manuscript AI declaration and responsible research reporting.

3. What AI Uses Journals Allow, Restrict, or Reject

The mere use of AI on its own does not justify rejection. Several improper uses of AI may cause significant issues in terms of publication and research integrity.

  • Fabrication of Information and References: The generative AI tool can generate false information and references. The citations, statements, statistics, and facts must be verified by the researcher. Elsevier explicitly states that AI-generated references could be incorrect or fabricated. Moreover, the falsified references could lead to manuscript rejection.
  • AI as a Replacement for Scientific Thinking: AI cannot replace the intellectual efforts of the author. If AI generates content for the manuscript while there is no analysis and interpretation of this content by the author, it leads to serious concerns about the originality of the work.
  • Undisclosed AI Assistance: The failure to disclose the AI assistance when the disclosure is required by the journal causes the issue of transparency and integrity. ICMJE states that undisclosed use of AI may lead to further steps and in some cases can be recognized as misconduct. Researchers should therefore understand the disclosure requirements journals impose before AI use manuscript submission.
  • Fabrication or Manipulation of Research Data: AI cannot be used for the fabrication or manipulation of the research data. This is a violation of research integrity.

4. AI Authorship Is Still Not Accepted

The most obvious present trend in the policies of major publishers is the fact that AI technology is not considered an author. AI technology cannot take responsibility for the accuracy and integrity of a manuscript, review the manuscript, answer any queries concerning the research, or be held accountable for any misconduct. Thus, researchers should continue to be responsible for the whole manuscript. According to ICMJE and Elsevier, AI technology should never be included in the list of authors or co-authors.

Springer Nature claims that LLM does not meet its requirements for authorship, though it allows using certain types of AI in copy editing, if it is only done to make text readable and to correct grammar, spelling, punctuation, and tone. This continuing AI authorship ban reinforces the principle of human accountability in academic publishing.

5. Confidentiality and Data Protection

Researchers must also think about the information that they provide to the AI system. Manuscripts that have not yet been published, patient information, confidential databases, research findings, and copyrighted information can pose problems for privacy, intellectual property, or confidentiality.

Researchers are advised by Elsevier that before using any AI tool, they should check the terms of service and the handling process of the information by the AI system. Editors

and reviewers have a tougher requirement since they must make sure that the manuscripts are kept confidential. ICMJE recommends that editors and reviewers do not upload manuscripts into an AI system where the information is not kept confidential.

This makes responsible AI use in research particularly important, especially when researchers use AI detection in academic writing or other AI-enabled systems involving manuscript content.

6. Current Trends in AI-Assisted Academic Publishing

There is a trend in the publication direction towards the adoption of responsible AI rather than banning AI from publications. [4]

Some of the current trends are:

  • Increased focus on AI disclosure statements.
  • Continuation of refusal to accept AI as an author.
  • Human authentication of AI-produced content.
  • Focus on AI-produced references and hallucinations.
  • Increased focus on confidentiality and intellectual property issues.
  • Policy division depending on the use of AI in writing, research, peer review, coding, and picture generation.
  • Increased need for documentation and reproducibility of AI usage.

These trends suggest that journals are increasingly focusing on AI governance and accountability instead of the mere question of AI usage by the author. The emphasis on responsible AI use in research is also influencing how researchers approach AI writing disclosure and manuscript preparation.

7. Practical Pre-Submission Checklist

Before submitting an AI-enhanced manuscript, researchers should:

  • Check the latest AI policy of the intended journal.
  • Detect all the substantive uses of the generative AI.
  • Validate references generated by AI and all factual statements.
  • Remove any invented or unsubstantiated data.
  • Ensure that the manuscript is consistent with the author’s original scientific interpretation.
  • Safeguard confidential and personal and unpublished data.
  • Add the mandatory AI disclosure statement.
  • Make sure that the human authors remain fully responsible for the manuscript.

Following these steps can help researchers meet AI use manuscript submission requirements and ensure compliance with the target journal’s journal AI policy.

Connect with us to explore how we can support you in maintaining academic integrity and enhancing the visibility of your research across the world!

Contact Us

Conclusion

The problem related to the usage of AI in academic publishing is not whether researchers will use artificial intelligence, but in how responsibly they will apply it. The editing made by humans that enhances language but does not replace the intellectual contribution of the author is becoming increasingly distinguished from content produced by AI and substituting for the work of the researcher.

This is why researchers need to see AI as a controlled helper rather than as a co-author itself. Disclosure of usage, verification, data safety, human accountability, and compliance with the policy of the journal are important for responsible publication. Thus, the best strategy is easy: use AI openly, verify all things, keep information safe, be accountable as a human, and follow the current policy of the journal.

Human Editing vs AI Writing Tools: What You Must Disclose and What Journals Still Reject. Our Pubrica consultants are here to guide you. [Get Expert Publishing Support] or [Schedule a Free Consultation]

Frequently Asked Questions (FAQs)

Choose a journal whose scope, target audience, article type, indexing, peer-review process, and publication requirements closely match your manuscript.

Q1 generally means that a journal is ranked in the top 25% of journals within a particular subject category and ranking system.

Verify the journal directly through the official Scopus Sources or Web of Science database rather than relying only on claims made on the journal’s website.

Check the journal’s ISSN, publisher, editorial board, peer-review process, indexing claims, publication fees, and publication ethics policies. Avoid relying solely on predatory-journal lists.

No. A Q1 ranking does not automatically mean the journal is the best fit. Scope, readership, indexing, peer review, publication ethics, fees, and relevance to your research should also be considered.

Compare journal scope, quartile, indexing, peer review, publication fees, editorial board, publication ethics, article types, and expected publication timelines.

References

  1. BaHammam A. S. (2025). The Transparency Paradox: Why Researchers Avoid Disclosing AI Assistance in Scientific Writing. Nature and science of sleep, 17, 2569–2574. https://doi.org/10.2147/NSS.S568375
  2. Cheng, A., Calhoun, A., & Reedy, G. (2025). Artificial intelligence-assisted academic writing: recommendations for ethical use. Advances in simulation (London, England), 10(1), 22. https://doi.org/10.1186/s41077-025-00350-6
  3. Phillips, J., & Wendler, D. (2014). Which alternatives should investigators disclose to research subjects?. The American journal of bioethics : AJOB, 14(4), 54–55. https://doi.org/10.1080/15265161.2014
  4. Oermann, M. H., Owens, J. K., Carter-Templeton, H., Peterson, G., & Bailey, H. E. (2025). Using Artificial Intelligence for Scholarly Writing. The American journal of nursing, 125(11), 52–55. https://doi.org/10.1097/AJN.0000