Comprehensive Summary
This cross-sectional study evaluated the rubric of the 54 MEDLINE-indexed PM&R journals surrounding policies around artificial intelligence use in manuscripts. Large language models likeChatGPT have become more commonly used in paper writing, increasing the need for regulations for authorship, plagiarism, and the reliability of sources. Slightly under half of PM&R journals have policies on AI use, and currently, no journals have explicit rules on AI usage. Journals with AI policies only allow AI to be used with manuscript editing with a declaration of use, and those that allow unlimited AI use still require authors to state the extent of AI use in the manuscript. Scientometric scores among journals with various AI policies were analyzed and determined to be indeterminate of journal impact factor. Overall, the study concludes that AI should be used with scientific integrity and that rapid technological evolution may require more established guidelines in the future.
Outcomes and Implications
The use of large language models (LLMs) have enabled medical researchers to use LLM(s) like ChatGPT for scientific publications, mainly manuscript drafting and editing. A medical and scientific implication of this advancement is upholding integrity as AI tools may not be accurate to results of data and do not have criteria for authorship. Additionally, data privacy leakage may also occur if AI privacy laws are not maintained, and AI tools may harmfully interpret or elaborate on data that may not be most essential to help patients and the overall community. More AI regulations may need to be created to ensure safety of privacy and intellect to maintain publication ethics in the medical and scientific community.