Exploring AI use policies in manuscript writing in cardiology and vascular journals

Background

Artificial intelligence (AI) technologies are rapidly evolving and offer efficiencies in manuscript generation however, this technology has raised concerns about the potential for bias, errors, and plagiarism to occur. In response, some journals have updated their author guidelines to address AI use.

Methods

We assessed author guidelines for 213 MEDLINE-indexed cardiovascular journals to evaluate policies on AI use in manuscript writing. Journal metrics such as CiteScore, Journal Impact Factor (JIF), Journal Citation Indicator (JCI), Source Normalized Impact per Paper (SNIP), and SCImago Journal Rank (SJR) were compared between journals with and without AI policies. We further analyzed the association between AI policy adoption and society affiliation. We reviewed the criteria for listing AI as an author and allowances for AI-generated content.

Results

Of 213 journals, 170 (79.8 %) had AI policies consistent across evaluations. Policies were present in 115 of 147 (78 %) cardiology journals and 113 of 127 (89 %) vascular journals. Furthermore, 111 of 143 (77.6 %) had AI-use policies, while 59 out of 70 (84.2 %) were unaffiliated journals. Journal metrics did not significantly differ between journals with and without AI policies (P > 0.05). Among journals with policies, 156 out of 158 (98.7 %) excluded AI as authors, while all allowed AI-assisted content.

Conclusion

Many cardiovascular journals address AI-generated content, but gaps remain in policies and disclosure requirements for AI-created manuscripts. The presence of AI-use policies was independent of journal metrics or society affiliation.