Artificial intelligence for manuscript writing: policies and implementation in cardiovascular journals
Brief Report

Artificial intelligence for manuscript writing: policies and implementation in cardiovascular journals

Todd A. Laffaye1 ORCID logo, Brian H. Carlson1, William K. Freeman2, Chadi Ayoub2

1Mayo Clinic Alix School of Medicine, Mayo Clinic, Phoenix, AZ, USA; 2Department of Cardiovascular Medicine, Mayo Clinic, Phoenix, AZ, USA

Correspondence to: Todd A. Laffaye, BA. Mayo Clinic Alix School of Medicine, Mayo Clinic, 5777 East Mayo Boulevard, Phoenix, AZ 85054, USA. Email: Laffaye.Todd@mayo.edu.

Abstract: Artificial intelligence (AI) has emerged as a widely used tool for writing, including in scientific research and publications. While its application to cardiovascular research is the focus of numerous studies, the policies related to its use for manuscript writing are rapidly evolving and not well understood. We sought to compare the policies of high-impact cardiovascular journals regarding AI for manuscript writing assistance and assess the prevalence of its use. Cardiovascular medicine journals with an SCImago Journal Rank (SJR) ≥3 and h-index ≥100 were screened for an AI policy. Journal policies were assessed for author disclosure requirements, standardization of disclosure section and language, and AI detection software used during the submission process. Each journal with an AI policy that required disclosure of its use was systematically searched to evaluate the prevalence of articles disclosing its use for writing assistance from January 2023 to August 2025. The number of publications with AI disclosure and publication characteristics was recorded. Seventeen journals met inclusion criteria and were screened for an AI policy, of which 14 journals (82%) contained such a policy. Among these, three journals (18%) had an AI policy that required disclosure, but that was not specific to AI use for manuscript writing. One journal (6%) did not require disclosure. The remaining three journals (18%) did not have any AI policy. None of the journals mandated a dedicated AI disclosure section or provided authors with standardized disclosure language. Fifteen journals (88%) used identifiable AI detection software, while only one posted this information publicly. Among the 14 journals with an AI disclosure policy, 11 AI-disclosing works were found. ChatGPT was the most common AI tool used (n=9, 82%). Journal policies regarding AI use for manuscript writing assistance vary widely, and therefore, there is a growing need for standardization. The prevalence of articles disclosing the use of AI was profoundly low across all journals evaluated, with significant variation in how AI use was disclosed. Having clear and consistent policies across journals and requiring authors to disclose their use of AI for manuscript writing is essential to uphold transparency and maintain medical research integrity.

Keywords: Artificial intelligence (AI); manuscript writing; disclosure policies; cardiovascular journals

Submitted Jul 08, 2025. Accepted for publication Sep 17, 2025. Published online Oct 24, 2025.

doi: 10.21037/cdt-2025-381

Artificial intelligence (AI) is an important and increasingly utilized tool throughout medicine (1). Within cardiovascular medicine, AI has been used to assist with clinical decisions, predict disease likelihood, and interpret imaging studies (2,3). As AI capabilities have grown, its use has extended to manuscript writing. Generative AI models can edit writing for spelling and grammar, enhance the quality and tone of human-written manuscripts, and create new content that is passable as manuscript-quality work (4). The International Committee of Medical Journal Editors (ICMJE) states in their April 2025 Recommendations for the Conduct, Reporting, Editing, and Publication of Scholarly Work in Medical Journals that journals should require authors to disclose the use of AI for any purpose related to the production of a submitted work, explicitly including writing assistance, in an appropriate section of the published work (5). Although the ICMJE and other international editorial and ethical guidelines commonly agree that AI tools cannot be listed as academic authors (6), individual journal and publisher policies for AI use in manuscript writing remain poorly understood.

While general AI policies within high-impact cardiovascular medicine journals have been reviewed (7), no prior study has systematically compared highly ranked cardiovascular journals for writing-related AI policies and assessed the prevalence of disclosure of AI use. This study sought to address this gap by reviewing editorial policies of high-impact cardiovascular medicine journals to identify policies, inconsistencies, and assess the publications in these journals in terms of prevalence and implementation regarding AI use for writing assistance.

Cardiovascular journals with a SCImago Journal Rank (SJR) ≥3 and h-index ≥100 were identified and screened for the existence of an AI policy. While many metrics of journal prestige exist, we selected the SJR because it is available to readers without cost, uses a prestige-weighted methodology (8), and includes more journals than other databases such as the Journal Citation Reports. We additionally required an h-index ≥100 to ensure long-term citation reliability (9). A search of each journal’s instructions for authors page was completed using publicly available statements from each journal’s website. Information extracted included: whether the journal possessed an AI policy, if the use of AI for writing assistance was required to be disclosed, recommended section for disclosure, specifications for disclosure language, whether the journal used AI detection software during the review process, and the publisher of each journal. We defined AI writing assistance as any application of AI related to drafting, editing, or proofreading. A journal was considered to have an AI policy if it contained any instructions for authors related to the use of AI.

For each journal that had an AI policy, the Wayback Machine from the Internet Archive (10) was applied to estimate when policies were created. An initial search for publications disclosing AI use was then performed within each journal’s respective search engine from January 1, 2023, to August 19, 2025, using a search string with Boolean connectors. All search fields and article types were included in the search. After this initial search, five additional terms were searched individually using the same methodology (Table 1).

Table 1

Search strategy summary

Items Specification
Date of search May 20, 2025 (initial search); August 19, 2025 (updated search)
Databases and other sources searched Journal AI policies were searched by visiting each journal’s instructions for authors or manuscript submission guidelines webpage. These webpages were accessed directly though each journal’s website. The Wayback Machine from Internet Archive was then applied to estimate when these policies were first created to inform the timeframe of the article search
For all journals that possessed an AI policy, a search was completed for articles disclosing the use of AI for manuscript writing assistance. This search was completed within each journal’s respective search engine
Search terms used The following search string containing multiple phrases in quotes separated by Boolean connectors was used to search for articles disclosing the use of AI: (“generative ai” OR “artificial intelligence” OR “language model” OR “ChatGPT”) AND (“declaration of generative” OR “disclosure” OR “writing assistance” OR “writing process” OR “during the preparation of this work”). This search string was inputted directly into each journal’s search bar on their respective websites
After an initial search using the search string above, five additional phrases in quotes were searched individually:
“the authors used ChatGPT”
“take full responsibility for the content”
“after using this tool/service”
“during the preparation of this work”
“during the course of preparing”
These terms were individually inputted directly into each journal’s search bar on their respective websites
If an advanced search feature was available, it was selected. The search string or search terms were inputted as shown into the primary search field. All search fields and article types were selected. The timeframe was entered, and the journal was re-selected in the advanced search if necessary. A manual full-text review of each article returned was completed
Timeframe AI policies were searched as of August 19, 2025
Articles disclosing the use of AI were searched between January 1, 2023, and August 19, 2025
Inclusion and exclusion criteria Cardiovascular journals with an SCImago Journal Rank ≥3 and h-index ≥100 were screened for the existence of an AI policy
All journals with an AI policy were evaluated for the prevalence of AI-disclosing works
Selection process Journals were selected by their SCImago Journal Rank and h-index as of May 16, 2025, as listed on the SCImago website

AI, artificial intelligence.

For each publication identified with an AI disclosure, information extracted included: journal, publication type, year of publication, disclosure section, if the disclosure section was congruent with the respective journal’s AI policy, and the AI program used and its stated purpose. The stated purpose was categorized as either drafting, editing, or proofreading, based on the description provided by the authors. The percentage of total published works that disclosed AI for writing assistance was calculated for each journal that had at least one AI disclosing publication. This was calculated by taking the total number of publications with AI disclosure for each journal and dividing it by the total number of articles returned from an empty search within each journal’s respective search engine between January 1, 2023, and August 19, 2025. The “Image” format was deselected for the European Heart Journal during this step to prevent duplication of results.

A sensitivity analysis was performed to confirm accuracy of findings. A random number generator was applied to select a random volume of each journal that was published within the study timeframe, and a manual review of identified publications was performed to recheck if there was disclosure of AI use for manuscript writing.

Seventeen journals met the initial inclusion criteria and were screened for an AI policy, with 14 journals (82%) found to possess such a policy. Among these, three journals (18%) had AI policies that required author disclosure but were not specific to AI use for manuscript writing. One journal (6%) had an AI policy that did not require author disclosure for manuscript writing. The remaining three journals (18%) did not have an AI policy and were excluded from the publication search. Of all 17 journals, 15 (88%) used identifiable AI detection software, while only the European Journal of Heart Failure (6%) posted this information publicly. Turnitin’s AI writing detection capability, which has been built into the iThenticate program (11), was the most common AI detection software tool used (n=15, 88%). The source of AI detection software appeared to be the publisher across all journals.

Of the 14 journals with AI policies, the most common disclosure section specified was the acknowledgments (n=9, 64%). Four journals (29%) specified two or more sections for AI disclosure. Notably, none of the journals assessed mandated a dedicated AI disclosure section or standardized disclosure language (Table 2).

Table 2

Cardiovascular journals with SJR ≥3 and h-index ≥100 and their associated AI manuscript writing policies, screening practices, and the prevalence of AI use

Journal Publisher AI writing policy (Y/N) Disclosure section specified by policy AI detection software Percent of total published works with an AI disclosure (January 1, 2023 to August 19, 2025), No. AI/No. total (%)
Nature Reviews Cardiology Springer Nature Yes Methods or other Geppetto, iThenticate 1/377 (0.265)
Journal of the American College of Cardiology Elsevier Yes AI disclosure not required iThenticate 1/18,025 (0.006)
Circulation LWW Yes Acknowledgements iThenticate 0
JAMA Cardiology AMA Yes Acknowledgements Unknown 0
European Journal of Heart Failure Wiley No N/A Papermill Detection, iThenticate N/A
JACC: Cardiovascular Imaging Elsevier Yes Acknowledgements iThenticate 0
JACC: Heart Failure Elsevier Yes Acknowledgements iThenticate 1/842 (0.119)
European Heart Journal OUP Yes† Methods or acknowledgements iThenticate 2/9,757 (0.020)
Circulation Research LWW Yes Acknowledgements iThenticate 0
Cardiovascular Research OUP No N/A iThenticate N/A
JACC: Cardiovascular Interventions Elsevier Yes Acknowledgements iThenticate 0
Circulation: Arrhythmia and Electrophysiology LWW Yes Acknowledgements iThenticate 0
Cardiovascular Diabetology Springer Nature Yes† Methods or other Geppetto, iThenticate 6/1,150 (0.522)
EuroIntervention Europa Digital & Publishing No N/A Unknown N/A
Europace OUP Yes† Methods or acknowledgements iThenticate 0
Circulation: Cardiovascular Interventions LWW Yes Acknowledgements iThenticate 0
Circulation: Heart Failure LWW Yes Acknowledgements iThenticate 0

†, general AI policy that was not specific for manuscript writing. AI, artificial intelligence; AMA, American Medical Association; JACC, Journal of the American College of Cardiology; LWW, Lippincott Williams & Wilkins; N/A, not available; OUP, Oxford University Press; SJR, SCImago Journal Rank.

After searching the 14 journals with an AI policy, 11 AI-disclosing publications were identified. Six (55%) were original research, two (18%) were cohort studies, one (9%) was an editorial piece, one (9%) was a discussion article, and one (9%) was a review article. The majority (n=6, 55%) were published in Cardiovascular Diabetology, two were published in the European Heart Journal (EHJ), and one was published in Nature Reviews Cardiology, the Journal of the American College of Cardiology (JACC), and JACC: Heart Failure, respectively. Nine works (82%) disclosed AI within the section specified by their respective journal’s policy.

Nine works (82%) used ChatGPT, one (9%) used the Wenxin Yiyan language model, and one (9%) used an unspecified large language model. The stated purpose of AI writing assistance included: drafting (n=1, 9%), editing of writing (n=7, 64%), and proofreading (n=3, 27%). Zero works were published in 2023, four (36%) were published in 2024, and seven (64%) were published in 2025. The sensitivity analysis with manual review of volumes identified no additional articles disclosing AI use for manuscript writing. Significant inconsistencies were identified across high-impact cardiovascular journals regarding AI writing policy and disclosure requirements. Additionally, we found that the current rates of AI disclosure were profoundly low across all journals evaluated. Among the policies reviewed, none required a dedicated AI disclosure section or specified standard disclosure language for authors. In our evaluation of journals with AI policies, only five out of 14 journals (36%) had published at least one article that contained an AI disclosure. Among these five was one journal with an AI policy that did not require disclosure for writing assistance but had published one article that disclosed the use of AI for this purpose.

Four of the 17 high-impact journals evaluated did not adhere to the ICMJE recommendations by either not providing authors with an AI policy or not requiring AI disclosure for manuscript writing. By comparing the characteristics and implementation of journal AI policies with observed rates of published disclosure, we offer the first comprehensive evaluation of AI use for manuscript writing assistance in the medical literature to date. There is currently a growing interest in using AI tools for manuscript drafting, editing, and formatting, as this can increase writing efficiency (12). The relative advantages and disadvantages of various AI tools for academic writing have already been described, as well as additional applications such as using AI for translation of non-English works (13,14). Despite their potential benefits, a clear understanding of the policies and prevalence of AI use for writing is essential, as AI tools can also commit errors and state factually incorrect information (15-17). These tools may also inadvertently reinforce harmful social biases because they reflect the data on which they were trained (18). Additionally, the highly automated nature of AI systems can make it difficult to control the quality of their work (19).

Based on these findings, we present several recommendations. First, we advise all journals to establish clear and consistent policies pertaining to AI use for writing assistance and outline these policies in their instructions for authors. Second, all journals should mandate a dedicated AI disclosure section and provide authors with a template of standardized disclosure language. Third, such a section should be present in all published works, attesting to the use of AI or lack thereof. Such measures will promote transparency and uphold scientific trust across cardiovascular medicine and the broader medical community.

This study should be interpreted in the context of several limitations. Our study was limited to high-impact journals, which provide a reasonable (and generally more rigorous) representation of the broader field of cardiovascular medicine, and it would be anticipated that lower-impact journals would have similar or even lower disclosure requirements and actual disclosure rates; future work is needed to evaluate the entire spectrum of journals, and to update our findings as the field evolves.

The highly variable nature of AI disclosure in its current form complicated the article search strategy for this study. The length of the search string used to identify publications with AI disclosure was limited to eight Boolean connectors by Science Direct’s search engine. Additionally, the content of all searches was limited by search engines universally ignoring stop words (common words in language that are typically filtered out by search processors) such as “the” and “used”. For example, this caused the phrase “the authors used ChatGPT” to be treated identically as “authors ChatGPT”, rendering the search ineffective. To ensure our search methodology remained robust, we introduced additional individually searched terms as described in the methods section. Finally, the profoundly low prevalence of articles disclosing the use of AI limited our analysis to descriptive statistics.

Policies regarding AI use for manuscript writing vary widely across high-impact cardiovascular journals. The prevalence of publications containing AI disclosures was very low and less than anticipated, with significant variation in how AI use was disclosed. It is likely that the use of AI for manuscript writing assistance is significantly under-reported across cardiovascular journals. The use of AI in this role should be reported, as it is recognized that these tools are not infallible, and informed human interpretation of scientific results remains essential. As AI applications become more integrated into academic medicine, there is a growing and urgent need to create standardized policies and disclosures for AI-assisted manuscript writing, to ensure the maintenance of integrity in scientific writing.

Acknowledgments

None.

Footnote

Peer Review File: Available at https://cdt.amegroups.com/article/view/10.21037/cdt-2025-381/prf

Funding: None.

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://cdt.amegroups.com/article/view/10.21037/cdt-2025-381/coif). C.A. serves as an unpaid editorial board member of Cardiovascular Diagnosis and Therapy from April 2025 to March 2027. C.A. is supported by the Clinician Engaged in Research award for his Mayo Clinic appointment; this did not fund this project. He is Chair of the AI Curriculum Committee, American Society of Echocardiography (unpaid). The other authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

  1. Xie Y, Zhai Y, Lu G. Evolution of artificial intelligence in healthcare: a 30-year bibliometric study. Front Med (Lausanne) 2024;11:1505692. [Crossref] [PubMed]
  2. Elias P, Jain SS, Poterucha T, et al. Artificial Intelligence for Cardiovascular Care-Part 1: Advances: JACC Review Topic of the Week. J Am Coll Cardiol 2024;83:2472-86. [Crossref] [PubMed]
  3. Itchhaporia D. Artificial intelligence in cardiology. Trends Cardiovasc Med 2022;32:34-41. [Crossref] [PubMed]
  4. Májovský M, Černý M, Kasal M, et al. Artificial Intelligence Can Generate Fraudulent but Authentic-Looking Scientific Medical Articles: Pandora's Box Has Been Opened. J Med Internet Res 2023;25:e46924. [Crossref] [PubMed]
  5. International Committee of Medical Journal Editors (ICMJE). Recommendations for the Conduct, Reporting, Editing, and Publication of Scholarly Work in Medical Journals [updated April 2025. Available online: https://www.icmje.org/recommendations/
  6. World Association of Medical Editors (WAME). Policies for Medical Journal Editors, prepared by the WAME Ethics and Policy Committee [updated May 2023]. Available online: https://www.wame.org/policies
  7. Inam M, Sheikh S, Minhas AMK, et al. A review of top cardiology and cardiovascular medicine journal guidelines regarding the use of generative artificial intelligence tools in scientific writing. Curr Probl Cardiol 2024;49:102387. [Crossref] [PubMed]
  8. González-Pereira B, Guerrero-Bote VP, Moya-Anegón F. A new approach to the metric of journals’ scientific prestige: The SJR indicator. Journal of Informetrics 2010;4:379-91.
  9. Mondal H, Deepak KK, Gupta M, et al. The h-Index: Understanding its predictors, significance, and criticism. J Family Med Prim Care 2023;12:2531-7. [Crossref] [PubMed]
  10. Internet Archive. Wayback Machine. Available online: https://web.archive.org/
  11. Turnitin LLC. AI writing detection in the new, enhanced Similarity Report view 2025. Available online: https://guides.ithenticate.com/hc/en-us/articles/27835303398541-AI-writing-detection-in-the-new-enhanced-Similarity-Report-view
  12. Maddali MM. Pro: Artificial Intelligence in Manuscript Writing: Advantages of Artificial Intelligence-Based Manuscript Writing to the Authors. Ann Card Anaesth 2025;28:198-200. [Crossref] [PubMed]
  13. Singh S, Kumar R, Maharshi V, et al. Harnessing Artificial Intelligence for Advancing Medical Manuscript Composition: Applications and Ethical Considerations. Cureus 2024;16:e71744. [Crossref] [PubMed]
  14. Kacena MA, Plotkin LI, Fehrenbacher JC. The Use of Artificial Intelligence in Writing Scientific Review Articles. Curr Osteoporos Rep 2024;22:115-21. [Crossref] [PubMed]
  15. Goodman RS, Patrinely JR, Stone CA Jr, et al. Accuracy and Reliability of Chatbot Responses to Physician Questions. JAMA Netw Open 2023;6:e2336483. [Crossref] [PubMed]
  16. Milutinovic S, Petrovic M, Begosh-Mayne D, et al. Evaluating Performance of ChatGPT on MKSAP Cardiology Board Review Questions. Int J Cardiol 2024;417:132576. [Crossref] [PubMed]
  17. Chen TC, Kaminski E, Koduri L, et al. Chat GPT as a Neuro-Score Calculator: Analysis of a Large Language Model's Performance on Various Neurological Exam Grading Scales. World Neurosurg 2023;179:e342-e347. [Crossref] [PubMed]
  18. Cohen JF, Moher D. Generative artificial intelligence and academic writing: friend or foe? J Clin Epidemiol 2025;179:111646. [Crossref] [PubMed]
  19. Pulido M. The challenge of artificial intelligence in medical writing and editing. Med Clin (Barc) 2024;163:186-8. [Crossref] [PubMed]
Cite this article as: Laffaye TA, Carlson BH, Freeman WK, Ayoub C. Artificial intelligence for manuscript writing: policies and implementation in cardiovascular journals. Cardiovasc Diagn Ther 2025;15(5):1107-1112. doi: 10.21037/cdt-2025-381

Article Options

Download Citation

Share