Skip to main content
To the Editor,
We found your commentary from the major family medicine journals about the use of Artificial Intelligence (AI) in family medicine writing and publication interesting(1).
We report a humorous experience with a family medicine journal that used AI for their copy-editing process. Our accepted article was about card studies, observational research in primary care(2) . Our study, currently in press, compared electronic collection of surveys with paper collection—meaning a survey on an 8x11 inch sheet of paper. The AI copy-editor misunderstood the context, converting the word “paper” to “article”, as in a manuscript or report, rather than a physical sheet of paper. The AI edited manuscript started with a history of card studies, “…researchers have utilized article card studies to assess primary care…” The AI changed the word “paper” card study to “article” card study throughout the abstract and manuscript, but not in the title, which created inconsistencies that could have been confusing to readers.
It took multiple readings by several authors during the review of the galley proofs to identify and understand the scope of the issue. The process of identifying and addressing these changes required significant time and effort to ensure the article accurately conveyed its findings. The journal editors confirmed that AI copy-editing was used and assured us the language would be reverted to its original meaning prior to publication.
This experience highlights the importance of authors conducting a full, meticulous review of their galley proofs—not just focusing on areas flagged by editors or areas of concern. Even seemingly straightforward or routine edits made by AI tools can lead to substantial misinterpretations or misrepresentations of the content, especially when dealing with nuanced terminology or specialized research topics.
Similarly, we urge journal editors to prioritize a final human review of AI-edited writing or editing to ensure that the published results align with the original intent and context(3). As AI tools become more integrated into the publication process, vigilance from both editors and authors is essential to uphold the integrity and clarity of scholarly work.
1) Schrager S, Seehusen DA, Sexton S, Richardson CR, Bowman MA, Rodríguez J, Morley C, Li L, James DD, Use of AI in Family Medicine Publications: A Joint Editorial From Journal Editors. The Annals of Family Medicine Jan 2025, 240575; DOI: 10.1370/afm.240575
2)Westfall JM, Zittleman L, Staton E, Parnes B, Smith PC, Niebauer L, Fernald D, Quintela J, Van Vorst R, Dickinson LM, Pace W. “Card Studies” - Observational Research in Practice-Based Research Networks. Annals of Family Medicine, January 2011;9: 63-68.
3)Liaw WR, Westfall JM, Williamson TS, Jabbarpour Y, Bazemore A. Primary Care: The Actual Intelligence Required for Artificial Intelligence to Advance Health Care and Improve Health. JMIR Med Inform. 2022 Mar 8;10(3):e27691. doi: 10.2196/27691. PMID: 35258464; PMCID: PMC8941433.