AI Disclosure Requirements in Research and Publishing
The tabbed box below shows the AI Policies of many major scholarly publishers.
Policy Reference Table
Federal Agencies
*Taken from Kennesaw State University AI Disclosure Requirements - Office of Research
|
Agency/Journal |
Specific Requirement |
AI as Author |
Direct URL |
|
NIH |
Peer reviewers prohibited from using AI for grant analysis; confidentiality violations if content uploaded to AI tools. |
Prohibited (implied) |
|
|
NSF |
Encourages AI disclosure in project descriptions; |
Not explicitly prohibited |
|
|
DOD |
General AI development guidelines; no specific research |
No explicit guidance |
Federal agency requirements establish the baseline
Federal funding agencies have established foundational requirements that researchers must follow regardless of journal policies. NIH leads with the most explicit restrictions, particularly prohibiting peer reviewers from using AI tools for grant application analysis. This June 2023 policy (NOT-OD-23-149) specifically bars natural language processors and large language models from peer review processes, citing confidentiality violations when grant content is uploaded to online AI tools.
NSF takes a more encouraging approach, suggesting researchers indicate AI use in project descriptions while prohibiting reviewers from using non-approved AI tools. The December 2023 guidance will be formalized in the 2025 Proposal and Award Policies and Procedures Guide. DOD notably lacks specific research disclosure requirements, focusing instead on operational AI development guidelines through their November 2023 AI Adoption Strategy.
Policy Reference Table
Major Publishers
*Taken from Kennesaw State University AI Disclosure Requirements - Office of Research
|
Agency/Journal |
Specific Requirement |
AI as Author |
Direct URL |
|
Elsevier |
Disclosure required except grammar checkers; no |
Prohibited |
|
|
Springer |
Copy editing exempt from disclosure; Methods section documentation for LLM use. |
Prohibited |
Major publishers establish cross-portfolio standards
Elsevier and Springer Nature, controlling thousands of journals across disciplines, have implemented comprehensive policies affecting researchers globally. Springer Nature distinguishes between AI-assisted copy editing (no disclosure required) and generative AI work (disclosure required), providing more nuanced guidance than most publishers.
Elsevier maintains stricter requirements, mandating disclosure for most AI use while prohibiting AI-generated images except in specific research contexts. Both publishers explicitly prohibit AI authorship and restrict reviewer AI use due to confidentiality concerns.
Policy Reference Table
High-Impact General Science
*Taken from Kennesaw State University AI Disclosure Requirements - Office of Research
|
Agency/Journal |
Specific Requirement |
AI as Author |
Direct URL |
|
Science |
Complete ban on AI-generated text; violation = scientific misconduct; full prompt disclosure required. |
Prohibited |
|
|
Nature |
LLM use documented in Methods; AI-generated images banned; copy editing exempt from disclosure |
Prohibited |
|
|
JAMA |
Detailed disclosure in Acknowledgments/Methods; |
Prohibited |
High-impact journals enforce strict boundaries
The most prestigious scientific journals maintain restrictive policies that effectively set industry standards. Science (AAAS) implements the harshest stance, completely banning AI-generated text and treating violations as scientific misconduct. Their policy requires full prompt disclosure in acknowledgments and methods sections, with AI tools explicitly prohibited from authorship.
Nature adopts a middle ground, prohibiting AI authorship and AI-generated images while allowing some AI assistance for copy editing without disclosure requirements. Their January 2023 policy specifically addresses peer review confidentiality, requiring reviewers to declare any AI use transparently.
JAMA demonstrates the most comprehensive disclosure framework with automated submission screening and detailed reporting requirements. Their March 2024 updated guidance includes specific institutional review board considerations for AI use in research design, representing the most thorough integration of AI oversight into the publication process.
Policy Reference Table
Engineering/Computer Science
*Taken from Kennesaw State University AI Disclosure Requirements - Office of Research
|
Agency/Journal |
Specific Requirement |
AI as Author |
Direct URL |
|
IEEE |
Disclosure in Acknowledgments with AI system |
Prohibited |
Disciplinary differences shape specific requirements
Engineering and computer science publications through IEEE maintain consistent policies across their extensive journal portfolio. Their April 2024 guidelines require acknowledgment section disclosure while explicitly prohibiting AI authorship and reviewer AI use. IEEE's approach emphasizes transparency while recognizing legitimate AI applications in technical fields.
Physical sciences publishers show variation: ACS requires detailed disclosure in acknowledgments with December 2024 updates providing specific guidance for AI-generated graphics, while APS limits AI to light editing only and completely prohibits AI-generated or modified images in Physical Review journals.
Life sciences publishers PLOS and Cell Press represent different philosophies. PLOS requires comprehensive disclosure in Methods sections with detailed evaluation descriptions, while Cell Press restricts AI to readability improvements using standardized disclosure templates.
Policy Reference Table
Physical Sciences
*Taken from Kennesaw State University AI Disclosure Requirements - Office of Research
|
Agency/Journal |
Specific Requirement |
AI as Author |
Direct URL |
|
ACS |
Disclosure in Acknowledgments/Methods; AI-generated |
Prohibited |
|
|
APS |
Light editing only; no AI-generated/modified images; disclosure to editors in cover letter. |
Prohibited |
Disciplinary differences shape specific requirements
Engineering and computer science publications through IEEE maintain consistent policies across their extensive journal portfolio. Their April 2024 guidelines require acknowledgment section disclosure while explicitly prohibiting AI authorship and reviewer AI use. IEEE's approach emphasizes transparency while recognizing legitimate AI applications in technical fields.
Physical sciences publishers show variation: ACS requires detailed disclosure in acknowledgments with December 2024 updates providing specific guidance for AI-generated graphics, while APS limits AI to light editing only and completely prohibits AI-generated or modified images in Physical Review journals.
Life sciences publishers PLOS and Cell Press represent different philosophies. PLOS requires comprehensive disclosure in Methods sections with detailed evaluation descriptions, while Cell Press restricts AI to readability improvements using standardized disclosure templates.
Policy Reference Table
Life Sciences
*Taken from Kennesaw State University AI Disclosure Requirements - Office of Research
|
Agency/Journal |
Specific Requirement |
AI as Author |
Direct URL |
|
PLOS |
Comprehensive disclosure in Methods/Acknowledgments; tool evaluation required. |
Prohibited |
|
|
Cell Press |
Readability/language improvement only; standardized disclosure template. |
Prohibited |
Disciplinary differences shape specific requirements
Engineering and computer science publications through IEEE maintain consistent policies across their extensive journal portfolio. Their April 2024 guidelines require acknowledgment section disclosure while explicitly prohibiting AI authorship and reviewer AI use. IEEE's approach emphasizes transparency while recognizing legitimate AI applications in technical fields.
Physical sciences publishers show variation: ACS requires detailed disclosure in acknowledgments with December 2024 updates providing specific guidance for AI-generated graphics, while APS limits AI to light editing only and completely prohibits AI-generated or modified images in Physical Review journals.
Life sciences publishers PLOS and Cell Press represent different philosophies. PLOS requires comprehensive disclosure in Methods sections with detailed evaluation descriptions, while Cell Press restricts AI to readability improvements using standardized disclosure templates.
Policy Reference Table
Humanities
*Taken from Kennesaw State University AI Disclosure Requirements - Office of Research
|
Agency/Journal |
Specific Requirement |
AI as Author |
Direct URL |
|
PMLA |
Full citation of AI-created content in manuscript at |
Prohibited |
|
|
Cambridge |
Clear explanation required; no plagiarism policy violations |
Prohibited |
Social sciences and humanities remain cautious
Humanities and social science journals demonstrate more conservative approaches, reflecting concerns about AI's ability to handle interpretive, contextual, and creative work. The American Journal of Political Science requires disclosure while discouraging AI use for substantial elements like literature reviews.
Cambridge University Press, the first major academic publisher to announce AI ethics policies in March 2023, prohibits AI authorship while requiring clear declaration of AI use. The Modern Language Association specifically addresses citation of AI tools while maintaining authorship restrictions.
Policy Reference Table
Social Sciences
*Taken from Kennesaw State University AI Disclosure Requirements - Office of Research
|
Agency/Journal |
Specific Requirement |
AI as Author |
Direct URL |
|
American Journal |
Disclosure in text/footnotes and author questionnaire; |
Prohibited |
|
|
American |
Brief submission disclosure for drafting/editing; thorough fact-checking required. |
Prohibited |
Social sciences and humanities remain cautious
Humanities and social science journals demonstrate more conservative approaches, reflecting concerns about AI's ability to handle interpretive, contextual, and creative work. The American Journal of Political Science requires disclosure while discouraging AI use for substantial elements like literature reviews.
Cambridge University Press, the first major academic publisher to announce AI ethics policies in March 2023, prohibits AI authorship while requiring clear declaration of AI use. The Modern Language Association specifically addresses citation of AI tools while maintaining authorship restrictions.
Key patterns and enforcement mechanisms
Several critical patterns emerge across all policies:
*Taken from Kennesaw State University AI Disclosure Requirements - Office of Research
Conclusion
The research landscape has rapidly adapted to AI integration with remarkably consistent core principles despite implementation variations. Researchers must navigate a complex matrix of federal requirements, journal policies, and disciplinary standards that universally prohibit AI authorship while requiring varying levels of disclosure transparency. The most restrictive policies from prestigious venues like Science and NIH are likely to influence broader adoption of conservative approaches, while more permissive frameworks may become standard for routine AI assistance in writing and analysis.
Guidance on these issues is still evolving, with policies likely to continue adapting as AI capabilities advance and research communities gain experience with appropriate integration boundaries.
This document was last updated by Dylan Goldblatt on July 30, 2025.
*Taken from Kennesaw State University AI Disclosure Requirements - Office of Research
